Forecasting: theory and practice

Fotios Petropoulos1,∗, Daniele Apiletti2 , Vassilios Assimakopoulos3 , Mohamed Zied Babai4 , Devon K. Barrow5 ,
Christoph Bergmeir6 , Ricardo J. Bessa7 , John E. Boylan8 , Jethro Browell9 , Claudio Carnevale10 , Jennifer L. Castle11 ,
Pasquale Cirillo12 , Michael P. Clements13 , Clara Cordeiro14,15 , Fernando Luiz Cyrino Oliveira16 , Shari De Baets17 ,
Alexander Dokumentov18 , Piotr Fiszeder19 , Philip Hans Franses20 , Michael Gilliland21 , M. Sinan Gönül22 ,
Paul Goodwin1 , Luigi Grossi23 , Yael Grushka-Cockayne24 , Mariangela Guidolin25 , Massimo Guidolin26 ,
Ulrich Gunter27 , Xiaojia Guo28 , Renato Guseo29 , Nigel Harvey30 , David F. Hendry31 , Ross Hollyman1 ,
Tim Januschowski32 , Jooyoung Jeon33 , Victor Richmond R. Jose34 , Yanfei Kang35 , Anne B. Koehler36 ,
Stephan Kolassa37,8 , Nikolaos Kourentzes38,8 , Sonia Leva39 , Feng Li40 , Konstantia Litsiou41 , Spyros Makridakis12 ,
Andrew B. Martinez42,43 , Sheik Meeran1 , Theodore Modis44 , Konstantinos Nikolopoulos45 , Dilek Önkal22 ,
Alessia Paccagnini46,47 , Ioannis Panapakidis48 , Jose M. Pavı́a49 , Manuela Pedio50,51 , Diego J. Pedregal52 ,
Pierre Pinson53 , Patrı́cia Ramos54 , David E. Rapach55 , J. James Reade56 , Bahman Rostami-Tabar57 ,
Michał Rubaszek58 , Georgios Sermpinis59 , Han Lin Shang60 , Evangelos Spiliotis3 , Aris A. Syntetos57 ,
Priyanga Dilini Talagala61 , Thiyanga S. Talagala62 , Len Tashman63 , Dimitrios Thomakos64 ,
Thordis Thorarinsdottir65 , Ezio Todini66,67 , Juan Ramón Trapero Arenas52 , Xiaoqian Wang35 , Robert L. Winkler68 ,
Alisa Yusupova8 , Florian Ziel69
1 School of Management, University of Bath, UK 2 Politecnico di Torino, Turin, Italy 3 Forecasting and Strategy Unit, School of Electrical and Computer
Engineering, National Technical University of Athens, Greece 4 Kedge Business School, France 5 Department of Management, Birmingham Business School,
University of Birmingham, UK 6 Faculty of Information Technology, Monash University, Melbourne, Australia 7 INESC TEC – Institute for Systems and
Computer Engineering, Technology and Science, Porto, Portugal 8 Centre for Marketing Analytics and Forecasting, Lancaster University Management School,
Lancaster University, UK 9 Department of Electronic and Electrical Engineering, University of Strathclyde, Glasgow, UK 10 Department of Mechanical
and Industrial Engineering, University of Brescia, Italy 11 Magdalen College, University of Oxford, UK 12 M Open Forecasting Center & Institute for the
Future, University of Nicosia, Nicosia, Cyprus 13 ICMA Centre, Henley Business School, University of Reading, UK 14 Faculdade de Ciências e Tecnologia,
Universidade do Algarve, Portugal 15 CEAUL, Faculdade de Ciências, Universidade de Lisboa, Portugal 16 Pontifical Catholic University of Rio de Janeiro
(PUC-Rio), Brazil 17 Department of Business Informatics and Operations Management, Faculty of Economics and Business Administration, Universiteit Gent,
Belgium 18 Let’s Forecast, Australia 19 Faculty of Economic Sciences and Management, Nicolaus Copernicus University in Torun, Poland 20 Econometric
Institute, Erasmus School of Economics, Rotterdam, The Netherlands 21 SAS, USA 22 Newcastle Business School, Northumbria University, Newcastle upon
Tyne, UK 23 Department of Economic Sciences, University of Verona, Italy 24 Darden School of Business, University of Virginia, USA 25 Department
of Statistical Sciences, University of Padua, Italy 26 Finance Department, Bocconi University and Baffi-CAREFIN Centre, Milan, Italy 27 Department of
Tourism and Service Management, MODUL University Vienna, Austria 28 UCL School of Management, University College London, UK 29 Department of
Statistical Sciences, University of Padua, Padova, Italy 30 Department of Experimental Psychology, University College London, UK 31 Nuffield College and
Institute for New Economic Thinking at the Oxford Martin School, University of Oxford, UK 32 Amazon Research, Germany 33 Korea Advanced Institute
of Science and Technology, South Korea 34 McDonough School of Business, Georgetown University, USA 35 School of Economics and Management, Beihang
University, Beijing, China 36 Miami University, Ohio, USA 37 SAP, Switzerland 38 Skövde Artificial Intelligence Lab, School of Informatics, University of
Skövde, Sweden 39 Department of Energy, Politecnico di Milano, Italy 40 School of Statistics and Mathematics, Central University of Finance and Economics,
Beijing, China 41 Manchester Metropolitan University Business School, UK 42 Office of Macroeconomic Analysis, US Department of the Treasury, Washington
DC, USA 43 GWU Research Program on Forecasting, Washington DC, USA 44 Growth Dynamics, Lugano, Switzerland 45 Durham University Business
School, Durham University, UK 46 Michael Smurfit Business School, University College Dublin, Ireland 47 Centre for Applied Macroeconomic Analysis,
Australia 48 Department of Electrical and Computer Engineering, University of Thessaly, Volos, Greece 49 GIPEyOP, UMMICS, Department of Applied
Economics, Universitat de Valencia, Spain 50 Accounting and Finance Department, University of Bristol, UK 51 Baffi Carefin Research Centre, Bocconi
University, Italy 52 ETSI Industrial, Universidad de Castilla-La Mancha, Ciudad Real, Spain 53 Technical University of Denmark, Kongens Lyngby, Denmark
54 Polytechnic of Porto, Porto, Portugal 55 Department of Economics, Saint Louis University, USA 56 Department of Economics, School of Politics, Economics
and International Relations, University of Reading, UK 57 Cardiff Business School, Cardiff University, UK 58 SGH Warsaw School of Economics, Collegium of
Economic Analysis, Poland 59 Adam Smith Business School, University of Glasgow, UK 60 Department of Actuarial Studies and Business Analytics, Macquarie
University, Australia 61 Department of Computational Mathematics, University of Moratuwa, Sri Lanka 62 Faculty of Applied Sciences, University of Sri
Jayewardenepura, Sri Lanka 63 Foresight, International Institute of Forecasters, USA 64 Department of Economics, University of Peloponnese, Tripolis, Greece
65 Norwegian Computing Center, Oslo, Norway 66 University of Bologna, Italy 67 Italian Hydrological Society, Bologna, Italy 68 Fuqua School of Business,
Duke University, Durham, USA 69 House of Energy Markets and Finance, University of Duisburg-Essen, Germany

∗ Corresponding author: [email protected]; [email protected]

Working paper, to be submitted to the International Journal of Forecasting October 26, 2020
 In theory, there is no difference between theory
and practice. But, in practice, there is.

Benjamin Brewster (1882)

Abstract

Forecasting has always been in the forefront of decision making and planning. The uncertainty that sur-
rounds the future is both exciting and challenging, with individuals and organisations seeking to minimise
risks and maximise utilities. The lack of a free-lunch theorem implies the need for a diverse set of fore-
casting methods to tackle an array of applications. This unique article provides a non-systematic review
of the theory and the practice of forecasting. We offer a wide range of theoretical, state-of-the-art mod-
els, methods, principles, and approaches to prepare, produce, organise, and evaluate forecasts. We then
demonstrate how such theoretical concepts are applied in a variety of real-life contexts, including oper-
ations, economics, finance, energy, environment, and social good. We do not claim that this review is an
exhaustive list of methods and applications. The list was compiled based on the expertise and interests
of the authors. However, we wish that our encyclopedic presentation will offer a point of reference for
the rich work that has been undertaken over the last decades, with some key insights for the future of the
forecasting theory and practice.
Keywords: review; encyclopedia; methods; applications; principles; time series; prediction.

2
Contents

1 Introduction 4

2 Theory 6
2.1 Introduction to forecasting theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 Pre-processing data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Statistical and econometric models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.4 Variable and model selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.5 Combining forecasts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.6 Data-driven methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.7 Methods for intermittent demands and count data . . . . . . . . . . . . . . . . . . . . . . . . 57
2.8 Reasoning and mining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
2.9 Forecasting by aggregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
2.10 Forecasting with judgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
2.11 Evaluation, validation, and calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
2.12 The future of forecasting theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

3 Practice 79
3.1 Introduction to forecasting practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.2 Operations and supply chain management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
3.3 Economics and finance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
3.4 Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
3.5 Environmental applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
3.6 Social good and demographic forecasting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
3.7 Systems and humans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
3.8 Other applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
3.9 The future of forecasting practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153

4 Forecasting: benefits, practices, value, and limitations 157

Appendix A List of acronyms 160

Appendix B Software 166

Appendix C Data sets 172

3
1. Introduction1

Forecasting has come a long way since early humans looked at the sky to see if the weather would be
suitable for hunting, and even since mid-19th-century hunters could get a forecast such as “a high of 40
with a chance of rain”. Now a hunter can look at a smartphone to instantly get hour-by-hour forecasts of
temperatures and probabilities of rain at multiple locations as well as videos of maps showing forecasted
weather patterns over the coming hours. Tailored forecasts of increasing sophistication can be generated
to inform important decisions of many different types by managers, public officials, investors, and other
decision makers.
Rapid advances in computing have enabled the analysis of larger and more complex data sets and
stimulated interest in analytics and data science. As a result, the forecaster’s toolbox of methods has
grown in size and sophistication. Computer science has led the way with methods such as neural networks
and other types of machine learning, which are getting a great deal of attention from forecasters and
decision makers. Other methods, including statistical methods such as Bayesian forecasting and complex
regression models, have also benefited from advances in computing. And improvements have not been
limited to those based on computing advances. For example, the literature on judgmental forecasting has
expanded considerably, driven largely by the “wisdom of crowds” notion.
The combining, or aggregation, of forecasts, which is not a new idea, has received increased attention in
the forecasting community recently and has been shown to perform well. For example, the top-performing
entries in the M4 forecasting competition run by Spyros Makridakis combined forecasts from multiple
methods. Many models have been developed to forecast the number of deaths that will be caused by
COVID-19, and combining the forecasts makes sense because it is hard to know which one will be the
most accurate. It is consistent with Bayesian ideas since it can be viewed as updating, with each individual
forecast added to the combined forecast (also called an ensemble) contributing some new information.
Despite the excitement surrounding these new developments, older methods such as exponential smooth-
ing and ARIMA are still valuable too. Exponential smoothing, along with other simple approaches like a
random walk, are quite robust and not as prone to overfitting as more complex methods. In that sense,
they are useful not only on their own merits, but as part of an ensemble that also includes more sophisti-
cated methods. Combined forecasts are more valuable if the forecasts come from methods that are diverse
so that their forecast errors are not highly correlated.
The conditions leading to larger, more sophisticated toolboxes for forecasters have also led to larger data
sets with denser grids and improved models in areas of application. This has happened with models of the
atmosphere, which are important in formulating improved weather forecasts. More detailed information
about customers and their preferences allows the development of improved models of customer behaviour
for managers. In turn, forecasting methods that can handle all of that information quickly are valuable for
decision-making purposes. This process has spurred an explosion in trying to gather information on the
internet.
In the Bayesian approach, inferences and forecasts are probabilistic in nature, and predictive proba-
bilities are forecasts. Risk is an important consideration in decision making, and probability forecasts can

1 This subsection was written by Robert L. Winkler.

4
quantify such risks. Theoretical work in probability forecasting has been active for some time, and decision
makers in many areas of practice have embraced the use of probability forecasts.
The U.S. National Weather Service began issuing probabilities of precipitation to the public in the
1960s. Yet extensive widespread use and dissemination of probabilities has only developed since the turn
of the century. Now probability forecasts are increasingly communicated to the public and used as inputs
in decision making. Nate Silver’s FiveThirtyEight.com report, gives probability forecasts for elections,
medicine and science, sporting events, economic measures, and many other areas, often looking at multi-
ple forecasting models individually and also combining them.
It is natural for people to desire certainty. When probability forecasts of precipitation were first dis-
seminated widely, many were very sceptical about them, with some accusing the forecasters of hedging
and saying “Don’t give me a probability. I want to know if it’s going to rain or not”. Of course, point
forecasts often are given along with probability forecasts. Frequent exposure to probabilities helps the
general public better understand, appreciate, and feel more comfortable with them. And the current situ-
ation in the world with COVID-19, increases in huge fires, big storms, political polarisation, international
conflicts, etc., should help them realise that we are living in an age with huge uncertainties, and forecasts
that quantify these uncertainties can be important. Where possible, visualisation can help, as indicated
by the saying that a picture is worth a thousand words. Examples are the cones of uncertainty on maps
in forecasts of the speed, severity, and future path of hurricanes, and the time line of the probability of a
team winning a game, updated quickly after each play.
Put simply, this is an exciting time for the field of forecasting. With all of the new theoretical develop-
ments and forecasting applications in practice, this article is very timely. Forecasting is so ubiquitous that
it is not possible to cover all of these developments in a single article. This article manages to cover quite
a few, and a good variety. Using short presentations for each one from an expert “close to the ground” on
that theoretical topic or field of practice works well to provide a picture of the current state of the art in
forecasting theory and practice.

5
2. Theory

2.1. Introduction to forecasting theory2

The theory of forecasting is based on the premise that current and past knowledge can be used to
make predictions about the future. In particular for time series, there is the belief that it is possible to
identify patterns in the historical values and successfully implement them in the process of predicting
future values. However, exact prediction of futures values is not expected. Instead, among the many
options for a forecast of a single time series at a future time period are an expected value (known as a
point forecast), a prediction interval, a percentile and an entire prediction distribution. This set of results
collectively could be considered to be “the forecast”. There are numerous other potential outcomes of a
forecasting process. The objective may be to forecast an event, such as equipment failure, and time series
may play only a small role in the forecasting process. Forecasting procedures are best when they relate
to a problem to be solved in practice. The theory can then be developed by understanding the essential
features of the problem. In turn, the theoretical results can lead to improved practice.
In this introduction, it is assumed that forecasting theories are developed as forecasting methods and
models. A forecasting method is defined here to be a predetermined sequence of steps that produces
forecasts at future time periods. Many forecasting methods, but definitely not all, have corresponding
stochastic models that produce the same point forecasts and can also be used to generate prediction in-
tervals and prediction distributions. A stochastic model makes assumptions about the process and the
associated probability distributions. Even when a forecasting method has an underlying stochastic model,
the model is not necessarily unique. For example, the simple exponential smoothing method has multi-
ple stochastic models, including state space models that may or may not be homoscedastic (i.e., possess
constant variance). The combining of forecasts from different methods has been shown to be a very suc-
cessful forecasting method. The combination of the corresponding stochastic models, if they exist, is itself
a model. Forecasts can be produced by a process that incorporates new and/or existing forecasting meth-
ods/models. Of course, these more complex processes would also be forecasting methods/models.
Consideration of the nature of the variables and their involvement in the forecasting process is essential.
In univariate forecasting, the forecasts are developed for a single time series by using the information from
the historical values of the time series itself. While in multivariate forecasting, other time series variables
are involved in producing the forecasts, as in time series regression. Both univariate and multivariate
forecasting may allow for interventions (e.g., special promotions, extreme weather). Relationships among
variables and other types of input could be linear or involve nonlinear structures (e.g., market penetration
of a new technology). When an explicit functional form is not available, methodologies such as simulation
or artificial neural networks might be employed. Theories from fields, such as economics, epidemiology,
and meteorology, can be an important part of developing these relationships. Multivariate forecasting
could, also, mean forecasting multiple variables simultaneously (e.g., econometric models).
The data or observed values for time series come in many different forms that may limit or determine
the choice of a forecasting method. In fact, there may be no historical observations at all for the item of in-
terest, when judgmental methods must be used (e.g., time taken to complete construction of a new airport).
The nature of the data may well require the development of a new forecasting method. The frequency of

2 This subsection was written by Anne B. Koehler.

6
observations can include all sorts of variations, such as every minute, hourly, weekly, monthly, and yearly
(e.g., the electricity industry needs to forecast demand loads at hourly intervals as well as long term de-
mand for ten or more years ahead). The data could be composed of everything from a single important
time series to billions of time series. Economic analysis often includes multiple variables, many of which
affect one another. Time series for businesses are likely to be important at many different levels (e.g., stock
keeping unit, common ingredients, or common size container) and, consequently, form a hierarchy of time
series. Some or many of the values might be zero; making the time series intermittent. The list of forms for
data is almost endless.
Prior to applying a forecasting method, the data may require pre-processing. There are basic details,
such as checking for accuracy and missing values. Other matters might precede the start of the fore-
casting method or be incorporated into the methods/models themselves. The treatment of seasonality is
such a case. Some forecasting method/models require de-seasonalised time series, while others address
seasonality within the methods/models. Making it less clear when seasonality is considered relative to a
forecasting method/model, some governmental statistical agencies produce forecasts to extend time series
into the future in the midst of estimating seasonal factors (i.e., X-12 ARIMA).
Finally, it is extremely important to evaluate the effectiveness of a forecasting method. The ultimate
application of the forecasts provides guidance in how to measure their accuracy. The focus is frequently
on the difference between the actual value and a point forecast for the value. Many loss functions have been
proposed to capture the “average” of these differences. Prediction intervals and percentiles can be used to
judge the value of a point forecast as part of the forecast. On the other hand, the quality of prediction
intervals and prediction distributions can themselves be evaluated by procedures and formulas that have
been developed (e.g., ones based on scoring rules). Another assessment tool is judging the forecasts by
metrics relevant to their usage (e.g., total costs or service levels).
In the remaining subsections of section 2, forecasting theory encompasses both stochastic modelling
and forecasting methods along with related aspects.

2.2. Pre-processing data

2.2.1. Unit root tests3
Time series models that are often used for forecasting are of the ARMA class, where ARMA is short
for autoregressive-moving average (Box et al., 2008). Before one can start with an iterative model-building
process (Franses et al., 2014), a forecaster needs to decide whether s/he creates models for the levels or the
differences of the data. Note that when the data are log-transformed, the differences of the logs approxi-
mate the growth rates.
The choice between growth rates and levels is guided by so-called unit root tests (Dickey and Fuller,
1979; Dickey and Pantula, 1987). Another way of phrasing it is that one needs to decide if the data have a
stochastic trend (if yes, then convert to growth rates) or a deterministic trend (stick to levels). Under the
null hypothesis of a stochastic trend, the data are non-stationary, meaning that they have no fixed mean
nor variance, and hence the test involves non-standard statistical theory (Phillips, 1987). There is also
the possibility to swap the hypotheses and take the levels as the null model (Kwiatkowski et al., 1992;

3 This subsection was written by Philip Hans Franses.

7
Hobijn et al., 2004). Non-stationary properties may also follow from structural breaks (Perron, 1989), so
it is reasonable to allow for potential breaks in the tests. The power of unit root tests is not large, and in
practice one often finds signals to consider growth rates (Nelson and Plosser, 1982). For forecasting, this is
not so much of a problem, as forecasts improve when growth rates are considered instead of levels (Franses
and Kleibergen, 1996).
For seasonal data, like quarters and months, one also needs to choose between levels and growth rates,
but now there are more choice options for the growth rates. For example for quarterly data, one can opt for
the current quarter minus the previous quarter (first differences with seasonal dummies, Miron, 1996), for
the current quarter minus the same quarter in the previous year (seasonal differences, Hylleberg, 1994), for
both at the same time (double differencing, Box et al., 2008), or for the current quarter minus a parameter
times the previous quarter (periodic differencing, Franses, 1994). In practice this means that one relies on
tests for a unit root like before and now also on tests for so-called seasonal unit roots (Hylleberg et al.,
1990; Franses, 1991; Ghysels et al., 1994) and/or on tests for periodic unit roots (Boswijk and Franses,
1996; Boswijk et al., 1997). Making the proper choice improves forecast accuracy (Kawasaki and Franses,
2004). Like structural breaks in non-seasonal data, it can happen that seasonal means shift abruptly, and
it is thus relevant to include these in the test regression (Franses and Vogelsang, 1998; Franses et al., 1997).
Allowing for mean shifts, if there are any, improves the out-of-sample forecasts (Paap et al., 1997).
Overall, the conclusion is that the proper choice for the type of growth rates is important for forecast
quality. Most tests are available in standard software, and asymptotic theory and critical values of the tests
are available for any seasonal format.

2.2.2. Time series decomposition4

Time series decomposition is an important building block for various forecasting approaches and a
crucial tools for statistical agencies. Seasonal decomposition is a way to present a time series as a function
of other time series, called components. Commonly used decompositions are additive and multiplicative,
where such functions are summation and multiplication correspondingly. If logs can be applied to time
series, any additive decomposition method can serve as multiplicative after applying log transformation to
the data.
The simplest additive decomposition of a time series with single seasonality comprises three compo-
nents: trend, seasonal component, and the “remainder”. It is assumed that the seasonal component has a
repeating pattern (thus sub-series corresponding to every season are smooth or even constant), the trend
component describes the smooth underlying mean and the remainder component is small and contains
noise.
The first attempt to decompose time series into trend and seasonality is dated to 1847 when Buys-Ballot
(1847) performed decomposition between trend and seasonality, modelling the trend by a polynomial and
the seasonality by dummy variables. Then, in 1884 Poynting (1884) proposed price averaging as a tool for
eliminating trend and seasonal fluctuations. Later, his approach was extended by Hooker (1901), Spencer
(1904) and Anderson and Nochmals (1914). Copeland (1915) was the first who attempted to extract the
seasonal component, and Macaulay (1931) proposed a method which is currently considered “classical”.

4 This subsection was written by Alexander Dokumentov.

8
 The main idea of this method comes from the observation that averaging a time series with window
size of the time series seasonal period leaves the trend almost intact, while effectively removes seasonal
and random components. At the next step, subtracting the estimated trend from the data and averaging
the result for every season gives the seasonal component. The rest becomes the remainder.
Classical decomposition led to a series of more complex decomposition methods such as X-11 (Shishkin
et al., 1967), X-11-ARIMA (Dagum, 1988; Ladiray and Quenneville, 2001), X-12-ARIMA (Findley et al.,
1998), and X-13-ARIMA-SEATS (Findley, 2005).
Seasonal trend decomposition using Loess (STL: Cleveland et al., 1990) takes iterative approach and
uses smoothing to obtain a better estimate of the trend and seasonal component at every iteration. Thus,
starting with an estimate of the trend component, the trend component is subtracted from the data, the
result is smoothed along sub-series corresponding to every season to obtain a “rough” estimate of the
seasonal component. Since it might contain some trend, it is averaged to extract this remaining trend,
which is then subtracted to get a detrended seasonal component. This detrended seasonal component is
subtracted from the data and the result is smoothed again to obtain a better estimate of the trend. This
cycle repeats a certain number of times.
Another big set of methods use a single underlining statistical model to perform decomposition. The
model allows computation of confidence and prediction intervals naturally, which is not common for itera-
tive and methods involving multiple models. The list of such methods includes TRAMO/SEATS procedure
(Monsell et al., 2003), the BATS and TBATS models (De Livera et al., 2011), various structural time series
model approaches (Harvey, 1990; Commandeur et al., 2011), and the recently developed seasonal-trend
decomposition based on regression (STR: Dokumentov, 2017; Dokumentov and Hyndman, 2018). The last
mentioned is one of the most generic decomposition methods allowing presence of missing values and
outliers, multiple seasonal and cyclic components, exogenous variables with constant, varying, seasonal
or cyclic influences, arbitrary complex seasonal schedules. By extending time series with a sequence of
missing values the method allows forecasting.

2.2.3. Anomaly detection and time series forecasting5

Temporal data are often subject to uncontrolled, unexpected interventions, from which various types
of anomalous observations are produced. Even though an anomaly is typically considered as an unusual,
sudden deviation from the norm (Hand, 2009), it is difficult to find a unified definition for an anomaly
and mostly application-specific (Unwin, 2019). Anomaly detection in forecasting literature has two main
focuses, which are conflicting in nature: one demands special attention be paid to anomalies as they can be
the main carriers of significant and often critical information such as fraud activities, disease outbreak, nat-
ural disasters, while the other down-grades the value of anomalies as it causes data quality issues (Talagala
et al., 2020a).
In time series and forecasting literature anomaly detection problems can be identified under three ma-
jor umbrella themes: detection of (i) contextual anomalies (point anomalies, additive anomalies) within a
given series, (i) anomalous sub-sequences within a given series, and (iii) anomalous series within a collec-
tion of series (Gupta et al., 2013; Talagala et al., 2020b). According to previous studies forecast intervals are

5 This subsection was written by Priyanga Dilini Talagala.

9
quite sensitive to contextual anomalies and the greatest impact on forecast are from anomalies occurring
at the forecast origin (Chen and Liu, 1993a).
The anomaly detection methods in forecasting applications can be categorised into two groups: (i)
model-based approaches and (ii) feature-based approaches. Model-based approaches compare the pre-
dicted values with the original data. If the deviations are beyond a certain threshold, the corresponding
observations are treated as anomalies (Luo et al., 2018b,a; Sobhani et al., 2020). Contextual anomalies and
anomalous sub-sequences are vastly covered by model-based approaches. Limitations in the detectabil-
ity of anomalous events depend on the input effects of external time series. Examples of such effects are
included in SARIMAX models for polynomial approaches. In nonlinear contexts an example is the gen-
eralised Bass model (Bass et al., 1994) for special life cycle time series with external control processes.
SARMAX with nonlinear perturbed mean trajectory as input variable may help separating the mean pro-
cess under control effects from anomalies in the residual process. Feature-based approaches, on the other
hand, do not rely on predictive models. Instead, they are based on the time series features measured using
different statistical operations that differentiate anomalous instances from typical behaviours (Fulcher and
Jones, 2014). Feature-based approaches are commonly used for detecting anomalous time series within a
large collection of time series. Under this approach, it first forecasts an anomalous threshold for the sys-
tems typical behaviour and new observations are identified as anomalies when they fall outside the bounds
of the established anomalous threshold (Talagala et al., 2019, 2020b). Most of the existing algorithms in-
volve a manual anomalous threshold. In contrast, Burridge and Robert Taylor (2006) and Talagala et al.
(2020b) use extreme value theory based data-driven anomalous thresholds. Approaches to the problem of
anomaly detection for temporal data can also be divided into two main scenarios: (i) batch processing and
(ii) data streams. The data stream scenario poses many additional challenges, due to nonstationarity, large
volume, high velocity, noisy signals, incomplete events and online support (Luo et al., 2018b; Talagala
et al., 2020b).
The performance evaluation of the anomaly detection frameworks is typically done using confusion
matrices (Luo et al., 2018b; Sobhani et al., 2020). However, these measures are not enough to evaluate the
performance of the classifiers in the presence of imbalanced data (Hossin and Sulaiman, 2015). Following
Ranawana and Palade (2006) and Talagala et al. (2019), Leigh et al. (2019) have used some additional
measures such as negative predictive value, positive predictive value and optimised precision to evaluate
the performance of their detection algorithms.

2.2.4. Exogenous variables and feature engineering6

Exogenous variables are those included in a forecasting system because they add value but are not be-
ing predicted themselves, and are sometimes called ‘features’. For example, a forecast of county’s energy
demand may be based on the recent history of demand (an endogenous variable), but also weather forecasts,
which are exogenous variables. Many time series methods have extensions that facilitate exogenous vari-
ables, such as autoregression with exogenous variables (ARX). However, it is often necessary to prepare
exogenous data before use, for example so that it matches the temporal resolution of the variable being
forecast (hourly, daily, and so on).

6 This subsection was written by Jethro Browell.

10
 Exogenous variables may be numeric or categorical, and may be numerous. Different types of predictor
present different issues depending on the predictive model being used. For instance, models based on
the variable’s absolute value can be sensitive to extreme values or skewness, whereas models based on the
variable value’s rank, such as tree-based models, are not. Exogenous variables that are correlated with one
another also poses a challenge for some models, and techniques such as regularisation and partial leased
squares have been developed to mitigate this.
Interactions between exogenous variables my also be important when making predictions. For exam-
ple, crop yields depend on both rainfall and sunlight: one without the other or both in excess will result
in low yields, but the right combination will result in high yields. Interactions may be included in linear
models by including the product of the two interacting exogenous variables as a feature in the model. This
is an example of feature engineering, the process of creating new features based on domain knowledge or
exploratory analysis of available data. In machine learning, many features may be created by combining ex-
ogenous variables speculatively and passed to a selection algorithm to identify those with predictive power.
Combinations are not limited to products, or only two interacting variables, and where many exogenous
variables are available, could include summary statistics (mean, standard deviation, range, quantiles...) of
groups of variables.
Where exogenous variables are numerous, dimension reduction may be applied to reduce the number
of features in a forecasting model. Dimension reduction transforms multivariate data into a lower dimen-
sional representation while retaining meaningful information about the original data. Principal compo-
nent analysis (PCA) is a widely used method for linear dimension reduction, and non-linear alternatives
are also available. In retail forecasting, for example, sales of thousands of products may be recorded but in-
cluding them all in a sales forecasting model may be impractical. Dimension reduction offers an alternative
to only using a subset of the available features.
Preparation of data for forecasting tasks is increasingly important as the volume of available data is
increasing in many application areas. Further details and practical examples can be found in Kuhn and
Johnson (2019) and Albon (2018) among other texts in this area. For deeper technical discussion and range
of non-linear dimension reduction algorithms, see Hastie et al. (2009).

2.3. Statistical and econometric models

2.3.1. Exponential smoothing models7
Exponential smoothing is one of the workhorses of business forecasting and despite the many advances
in the field, ETS is always a tough benchmark to bear in mind. The development of exponential smoothing
dates back to 1944, where Robert G. Brown through a mechanical computing device estimated key vari-
ables for fire-control on the location of submarines (Gardner, 1985). More details about the state of the art
of exponential smoothing can be found in Gardner (2006).
The idea behind exponential smoothing relies on the weighted average of past observations, where
that weight decreases exponentially as one moves away from the present observations. The appropriate
exponential smoothing method depends on the components that appear in the time series. For instance,
in case that no clear trend or seasonal pattern is present, the simplest form of exponential smoothing

7 This subsection was written by Juan Ramón Trapero Arenas.

11
methods known as Simple (or Single) Exponential Smoothing (SES) is adequate. In some references, is also
known as Exponentially Weighted Moving Average (Harvey, 1990). The formula for SES can be obtained
from minimising the discounted least squares error function and expressing the resulting equation in a
recursive form (Harvey, 1990). If observations do not have the same weight, the ordinary least squares
cannot be applied. On the other hand, the recursive form is very well-suited for saving data storage.
In order to use SES, we need to initialise the algorithm and to define the exponential smoothing param-
eter. Those values can be obtained by minimising the sum of squares of the one-step ahead forecast errors.
Traditionally, the initialisation is done by using either ad hoc values or a heuristic scheme (Hyndman et al.,
2008). The estimation of the smoothing parameter usually is restricted to values between 0 and 1.
Once SES is defined, the method only provides point forecasts, i.e., forecasts of the mean. Nonetheless,
for SES as well as the rest of exponential smoothing methods, it is of vital importance for many applica-
tions to provide density (probabilistic) forecasts. To that end, Hyndman et al. (2002) extended exponential
smoothing methods under state space models to equip them with a statistical framework capable of pro-
viding future probability distributions. Note the difference between traditional exponential smoothing
methods and exponential smoothing models (under the state space approach). The former only provide
point forecasts, meanwhile the latter also offers probabilistic forecasts, which obviously includes predic-
tion intervals.
So far, we have introduced the main concepts using SES, however, real time series can include other
components as trends, seasonal patterns, cycles, and the irregular (error) component. Fortunately, for
many combinations of components a particular exponential smoothing can be chosen. Pegels (1969) pro-
posed a first classification of exponential smoothing methods, later extended by Gardner (1985) and Tay-
lor (2003a). The state space framework mentioned above, developed by Hyndman et al. (2002), allowed
to compute the likelihood for each exponential smoothing model and, thus, model selection criteria such
as AIC could be used to automatically identify the appropriate exponential smoothing model. Note that
the equivalent state space formulation was derived by using a single source of error instead of a mul-
tiple source of error (Harvey, 1990). Hyndman et al. (2008) utilised the notation (E,T,S) to classify the
exponential smoothing models, where those letters refer to the following components: Error, Trend, and
Seasonality. This notation has gained popularity because the widely-used forecast package (Hyndman et al.,
2020) for R statistical software, and nowadays exponential smoothing is frequently called ETS.

2.3.2. Time-series regression models8

The key idea of linear regression models is that a target (or dependent, forecast, explained, regress)
variable, i.e., a time series of interest, can be forecast through other regressor (or independent, predic-
tor, explanatory) variables, i.e., time series or features, assuming that a linear relationship exists between
them. For example, daily product sales may be forecast using information related with past sales, prices,
advertising, promotions, special days, and holidays.
In order to estimate the model, forecasters typically minimise the sum of the squared errors (least
squares estimation), using the observations available for fitting the model to the data (Ord et al., 2017).
If the model is simple, consisting of a single regressor, then two coefficients are computed, which are the

8 This subsection was written by Vassilios Assimakopoulos.

12
coefficient of the regressor and a constant. When more regressor variables are considered, the model is
characterised as a multiple regression one and additional coefficients are estimated.
A common way to evaluate how well a linear regression model fits the data, is through the coefficient of
determination, indicating the proportion of variation in the target variable explained by the model. Values
close to one indicate sufficient goodness-of-fit, while values close to zero insufficient fitting. However,
goodness-of-fit should not be confused with forecastability (Harrell, 2015). When the complexity of the
model is increased, i.e., more regressors are considered, the value of the coefficient will also rise, even
if such additions lead to overfitting. Thus, regression models should be evaluated using cross-validation
approaches, approximating the post-sample accuracy of the model, or measures that account for model
complexity, such as information criteria and the adjusted coefficient of determination (James et al., 2013).
Other diagnostics are the standard deviation of the residuals and the t-values of the regressors. Residual
standard error summarises the average error produced by the model given the number of regressors used,
thus accounting for overfitting. The t-values measure the impact of excluding regressors from the model
in terms of error, given the variation in the data, thus highlighting the importance of the regressors.
To make sure that the produced forecasts are reliable, the main assumptions made by the model must
be fulfilled. This means that (i) the residuals must be normally distributed with an average value of zero,
(ii) the residuals must display insignificant autocorrelation, (iii) the variability of the residuals should be
equal across time (no heteroscedasticity present), and (iv) the correlation between the residuals and the
observations of the regressor must be zero. If these assumptions are violated, that may mean either that
part of the variance of the target variable has not been explained by the model (other or more regressors are
needed), or that outliers and influential observations have negatively affected the estimation of the model.
Apart from time series regressors, regression models can also exploit categorical (dummy or indicator)
variables (Hyndman and Athanasopoulos, 2018) which may e.g., inform the model about promotions, spe-
cial events, and holidays (binary variables), the day of the week or month of the year (seasonal dummy
variables provided as one-hot encoded vectors), trends and structural changes, and the number of trad-
ing/working days included in the examined period. In cases where the target series is long and displays
complex seasonal patterns, additional regressors such as Fourier series and lagged values of both the target
and the regressor variables may become useful.

2.3.3. Theta method and models9

In the age of vast computing power and computational intelligence, the contribution of simple fore-
casting methods is possibly not en vogue; the implementation of complicated forecasting systems becomes
not only expedient but possibly desirable. Nevertheless forecasting, being a tricky business, does not
always favour the complicated or the computationally intensive. Enter the theta method. From its be-
ginnings 20 years back in Assimakopoulos and Nikolopoulos (2000) to recent advances in the monograph
of Nikolopoulos and Thomakos (2019), to other work in-between and recently too, the theta method has
emerged as not only a powerfully simple but also enduring method in modern time series forecasting. The
original idea has been now fully explained and understood and, as Nikolopoulos and Thomakos (2019)
have shown, even the reverend AR(1) model forecast is indeed a theta forecast – and it has already been
shown by Hyndman and Billah (2003) that the theta method can represent simple exponential smoothing

9 This subsection was written by Dimitrios Thomakos.

13
(SES) with drift forecasts as well. In its simplest form the method generates a forecast from a linear com-
bination of the last observation and some form of “trend” function, be that a constant, a linear trend, a
non-parametric trend or a non-linear trend. The weights on this linear combination, the number of trend
functions and their nature and other aspects on expanding the method have been recently researched ex-
tensively.
The main literature has two strands. The first one details the probabilistic background of the method
and derives certain theoretical properties, as in Hyndman and Billah (2003), Thomakos and Nikolopoulos
(2012, 2015) and a number of new theoretical results in Nikolopoulos and Thomakos (2019). The work of
Thomakos and Nikolopoulos provided a complete analysis of the theta method under the unit root data
generating process, explained its success in the M3 competition (Makridakis and Hibon, 2000), introduced
the multivariate theta method and related it to cointegration and provided a number of other analytical
results for different trend functions and multivariate forecasting. The second strand of the literature ex-
pands and details various implementation (including hybrid approaches) of the method, as in the theta
approach in supply chain planning of Nikolopoulos et al. (2012), the optimised theta models and their
relationship with state space models in Fioruci et al. (2015) and Fiorucci et al. (2016), hybrid approaches
as in Theodosiou (2011) and Spiliotis et al. (2019a), to the very latest generalised theta method of Spiliotis
et al. (2020a). These are major methodological references in the field, in addition to many others of pure
application of the method.
The theta method is also part of the family of adaptive models/methods, and a simple example il-
lustrates the point: the AR(1) forecast or the SES forecast are both theta forecasts but they are also both
adaptive learning forecasts, as in the definitions of the recent work by Kyriazi et al. (2019). As such, the
theta forecasts contain the basic building blocks of successful forecasts: simplicity, theoretical foundations,
adaptability and performance enhancements. Further research on the usage of the theta method within
the context of adaptive learning appears to be a natural next step.
Given the simplicity of its application, the freely available libraries of its computation, its scalability
and performance, the theta method should be considered as a critical benchmark henceforth in the litera-
ture – no amount of complexity is worth its weight if it cannot beat a single Greek letter!

2.3.4. Autoregressive integrated moving average (ARIMA) models10

The fundamental planks of a forecasting task are collecting the required data and finding a pattern in
the data. Then comes the job of identifying a model to represent the pattern, i.e., fitting the model to the
data. The procedure has been the staple of forecasters in many of the methods they used. Sometimes when
the pattern is not conspicuous, it becomes a challenge. A wise next step would be exploring the data for
any hidden patterns. Box et al. (1976) did one such exploration with regards to time series data. They
realised that although there may not be any visible patterns in some data series, a pattern could emerge
if they correlated the series with itself, i.e., by finding autocorrelations. As an exemplar for finding an
autocorrelation, one can offset the data by one period (i.e., making the series to lag by one period) to get
a different series and then find the correlation between these — which results in an autocorrelation of
lag 1. If the second series is obtained by offsetting the original series by two periods, then one gets an
autocorrelation of lag 2. This process continues for other lags.

10 This subsection was written by Sheik Meeran.

14
 Box et al. (1976) made use of this ‘autocorrelation phenomenon’ and incorporated the same into a
self-regressing model, i.e., an autoregressive model, represented by the AR part of ARIMA. They further
expanded the idea of extracting hidden patterns by finding a relationship between a data in one period
and the error in the previous period, and thus came the Moving Average model, i.e., the MA portion of
ARIMA. Curiosity about the name ‘Moving Average’ is palpable among forecasting enthusiasts. Some
think the name is arbitrary while others think the history behind this name is lost. Yet another view is
“each value of yt can be thought of as a weighted moving average of the past few forecasts” (Hyndman and
Athanasopoulos, 2018), where y is the forecast variable.
Sometimes, the pattern in the data could be even harder to spot as the pattern is ‘Integrated’ (sub-
merged) with the trend in the data. Such data are called non-stationary. Hence a need for differencing the
series to remove the trend and extract the pattern. ‘I’ in ARIMA denotes the integrated nature of the data.
There are alternative views on this, namely ‘integrated’ could mean ‘all-inclusive’. Some consider it could
have come from the need for the differenced data to be integrated to get the values at the same level as in
the original series.
ARIMA has three parameters, p, d, and q, indicating the number of AR terms, the number of differenc-
ing, and the number of MA terms respectively. An ARIMA model is denoted as ARIMA(p, d, q). The data
series could have, besides, seasonal patterns that could overlap with the non-seasonal pattern that is seen
above. Such models are called Seasonal ARIMA (SARIMA) models represented by ARIMA(p, d, q)(P , D, Q)s ,
where P , D, and Q are the seasonal parameters and the s is the periodicity. Enumerating such models could
be quite demanding. Some software use autocorrelation function (ACF) and partial autocorrelation func-
tion (PACF) to identify suitable models. The function auto.arima of the forecast package (Hyndman et al.,
2020) for R statistical software compares models using information criteria, and has been found to be very
effective and increasingly being used in ARIMA modelling.
A popular extension to ARIMA models is called ARIMAX, implemented by incorporating additional
exogenous variables (regressors) that are external to and different from the forecast variable.

2.3.5. ARCH/GARCH models11

Volatility has been recognised as a primary measure of risks and uncertainties (Markowitz, 1952;
Sharpe, 1964; Taylor et al., 2009; Gneiting, 2011a). Estimating future volatility for measuring the un-
certainty of forecasts is imperative for probabilistic forecasting. Yet, the right period in which to estimate
future volatility has been controversial as volatility based on too long a period will make irrelevant the fore-
cast horizon of our interests, whereas too short a period results in too much noise (Engle, 2004). An alterna-
tive to this issue is the dynamic volatility estimated through the autoregressive conditional heteroscedas-
ticity (ARCH) proposed by Engle (1982), and the generalised autoregressive conditional heteroscedasticity
(GARCH) model proposed by Bollerslev (1987). The ARCH model uses the weighted average of the past
squared forecast error whereas the GARCH model generalises the ARCH model by further adopting past
squared conditional volatilities. The GARCH model is the combination of (i) a constant volatility, which
estimates the long-run average, (ii) the volatility forecast(s) in the last steps, and (iii) the new information
collected in the last steps. The weightings on these components are typically estimated with maximum
likelihood. The models assume a residual distribution allowing for producing density forecasts. One of

11 This subsection was written by Jooyoung Jeon.

15
the benefits of the GARCH model is that it can model heteroscedasticity, the volatility clustering character-
istics of time series (Mandelbrot, 1963), a phenomenon common to many time series where uncertainties
are predominant. Volatility clustering comes about as new information tends to arrive time clustered and
a certain time interval is required for the time series to be stabilised as the new information is initially
recognised as a shock.
The GARCH model has been extended in the diverse aspects of non-linearity, asymmetry and long
memory. Among many such extensions, the Exponential GARCH (EGARCH) model by Nelson (1991) uses
log transformation to prevent negative variance; the Threshold GARCH (TGARCH) model by Zakoian
(1994) allows for different responses on positive and negative shocks. A small piece of information can have
more impact when the time series is under stress than under a stable time series (Engle, 2004). Another
pattern often observed in the volatility time series is slowly decaying autocorrelation, also known as a
long memory pattern, which Baillie et al. (1996) capture using a slow hyperbolic rate of decay for the
ARCH terms in the fractionally integrated GARCH (FIGARCH) model. Separately, in a further approach
to directly estimating long term volatility, the GARCH-MIDAS (Mixed Data Sampling) model proposed by
Engle et al. (2013) decomposes the conditional volatility into the short-term volatility, as captured by the
traditional GARCH, and the long-term volatility represented by the realised volatilities.
The univariate GARCH models surveyed so far have been exended to multivariate versions, in order to
model changes in the conditional covariance in multiple time series, resulting in such examples as the VEC
(Bollerslev, 1987) and BEKK (Engle and Kroner, 1995), an acronym derived from Baba, Engle, Kraft, and
Kroner. The VEC model, a direct generalisation of the univariate GARCH, requires more parameters in
the covariane matrices and provides better fitness at the expense of higher estimation costs than the BEKK.
The VEC model has to ensure the positivity of the covariance matrix with further constraints, whereas the
BEKK model and its specific forms, e.g., factor models, avoid this positivity issue directly at the model
specification stage. In an effort to further reduce the number of parameters to be estimated, the linear and
non-linear combinations of the univariate GARCH models, such as the constant conditional correlation
model of Bollerslev (1990) and the dynamic conditional correlation models of Tse and Tsui (2002) and of
Engle (2002), were investigated.

2.3.6. Forecasting for multiple seasonal cycles12

With the advances in digital data technologies, data is recorded more frequently in many sectors such
as energy (Wang et al., 2016), healthcare (Whitt and Zhang, 2019), transportation (Gould et al., 2008),
and telecommunication (Meade and Islam, 2015a). This often results in time series that exhibit multiple
seasonal cycles (MSC) of different lengths. Forecasting problems involving such series have been increas-
ingly drawing the attention of both researchers and practitioners leading to the development of several
approaches.
Multiple Linear Regression (MLR) is a common approach to model series with MSC (Kamisan et al.,
2018; Rostami-Tabar and Ziel, 2020). While MLR is fast, flexible, and uses exogenous regressors, it does
not allow to decompose components and change them over time. Building on the foundation of the regres-
sion, Facebook introduced Prophet (Taylor and Letham, 2018), an automated approach that utilises the
Generalised Additive Model (Hastie and Tibshirani, 1990). Although the implementation of Prophet may

12 This subsection was written by Bahman Rostami-Tabar.

16
be less flexible, it is easy to use, robust to missing values and structural changes, and can handles outliers
well.
Some studies have extended the classical ARIMA and Exponential Smoothing (ETS) methods to accom-
modate MSC. Multiple/multiplicative Seasonal ARIMA (MSARIMA) model is an extension of ARIMA for
the case of MSC (Taylor, 2003b). MSARIMA allows for exogenous regressors and terms can evolve over
time, however, it is not flexible, and the computational time is high. Svetunkov and Boylan (2020) intro-
duced the Several Seasonalities ARIMA (SSARIMA) model which constructs ARIMA in a state-space form
with several seasonalities. While SSARIMA is flexible and allows for exogenous regressors, it is computa-
tionally expensive, especially for high frequency series.
Taylor (2003b) introduced Double Seasonal Holt-Winters (DSHW) to extend ETS for modelling daily
and weekly seasonal cycles. Following that, Taylor (2010) proposed a triple seasonal model to consider the
intraday, intraweek and intrayear seasonalities. Gould et al. (2008) and Taylor and Snyder (2012) instead
proposed an approach that combines a parsimonious representation of the seasonal states up to a weekly
period in an innovation state space model. With these models, components can change, and decomposition
is possible. However, the implementation is not flexible, the use of exogenous regressors is not supported,
and the computational time could be high.
An alternative approach for forecasting series with MSC is TBATS (De Livera et al., 2011). TBATS
uses a combination of Fourier terms with an exponential smoothing state space model and a Box-Cox
transformation, in an entirely automated manner. It allows for terms to evolve over time and produce
accurate forecasts. Some drawbacks of TBATS, however, are that it is not flexible, can be slow, and does not
allow for covariates.
In response to shortcomings in current models, Forecasting with Additive Switching of Seasonality,
Trend and Exogenous Regressors (FASSTER) has been proposed by O’Hara-Wild and Hyndman (2020).
FASSTER is fast, flexible and support the use of exogenous regressors into a state space model. It extends
state space models such as TBATS by introducing a switching component to the measurement equation
which captures groups of irregular multiple seasonality by switching between states.
In recent years, Machine Learning (ML) approaches have also been recommended for forecasting time
series with MSC. MultiLayer Perceptron (MLP: Dudek, 2013; Zhang and Qi, 2005), Recurrent Neural Net-
works (RNN: Lai et al., 2018), Generalised Regression Neural Network (GRNN: Dudek, 2015), and Long
Short-Term Memory Networks (LSTM Zheng et al., 2017) have been applied on real data (Bandara et al.,
2020a; Xie and Ding, 2020) with promising results. These approaches are flexible, allow for any exogenous
regressor and suitable when non-linearity exists in series, however interpretability might be an issue for
users (Makridakis et al., 2018).

2.3.7. State-space models13

State Space (SS) systems are a very powerful and useful framework for time series and econometric
modelling and forecasting. Such systems were initially developed by engineers, but have been widely
adopted and developed in Economics as well (Harvey, 1990; Durbin and Koopman, 2012). The main
distinguishing feature of SS systems is that the model is formulated in terms of states, which are a set

13 This subsection was written by Diego J. Pedregal.

17
of variables usually unobserved, but which have some meaning. Typical examples are trends, seasonal
components or time varying parameters.
A SS system is built as the combination of two sets of equations: (i) state or transition equations which
describe the dynamic law governing the states between two adjacent points in time, and (ii) observation
equations which specify the relation between observed data (both inputs and outputs) and the unobserved
states. Once a SS system is fully specified, the core problem is to provide optimal estimates of states and
their covariance matrix over time. This can be done in two ways, either by looking back in time using the
well-known Kalman filter (useful for online applications) or taking into account the whole sample provided
by smoothing algorithms (typical of offline applications) (Anderson and Moore, 1979).
Given any set of data and a specific model, the system is not fully specified in most cases because
it usually depends on unknown parameters scattered throughout the system matrices that define the SS
equations. Estimation of such parameters is normally carried out by Maximum Likelihood defined by
prediction error decomposition (Harvey, 1990).
Non-linear and non-Gaussian models are also possible, but at the cost of a higher computational burden
because more sophisticated recursive algorithms have to be applied, like the extended Kalman filters and
smoothers of different orders, particle filters (Doucet and Gordon, 2001), Unscented Kalman filter and
smoother (Julier and Uhlmann, 1997), or simulation of many kinds, like Monte Carlo, bootstrapping or
importance sampling (Durbin and Koopman, 2012).
The paramount advantage of SS systems is that they are not a particular model or family of models
strictly speaking, but a container in which many very different model families may be implemented, indeed
many treated in other sections of this paper. The following is a list of possibilities, not at all exhaustive:

• Univariate models with or without inputs: regression, ARIMAX, transfer functions, non-multiplicative
exponential smoothing, structural unobserved components, Hodrick-Prescott filter, spline smooth-
ing.
• Fully multivariate: natural extensions of the previous ones plus echelon-form VARIMAX, Structural
VAR, VECM, Dynamic Factor models, panel data.
• Non-linear and non-Gaussian: TAR, ARCH, GARCH, Stochastic Volatility (Durbin and Koopman,
2012), Dynamic Conditional Score (Harvey, 2013), Generalised Autoregressive Score (Creal et al.,
2013), multiplicative exponential smoothing, multiplicative unobserved components.
• Other: periodic cubic splines, periodic unobserved components models, state dependent models,
Gegenbauer long memory processes (Dissanayake et al., 2018).

Once any researcher or practitioner becomes acquainted to a certain degree with the SS technology,
some important advanced issues in time series forecasting may be comfortably addressed (Casals et al.,
2016). It is the case, for example, of systems block concatenation, systems nesting in errors or in variables,
treating errors in variables, continuous time models, time irregularly spaced data, mixed frequency mod-
els, time varying parameters, time aggregation, hierarchical and group forecasting (time, longitudinal or
both: Villegas and Pedregal, 2018), homogeneity of multivariate models (proportional covariance structure
among perturbations), etc.
All in all, the SS systems offer a framework capable of handling many modelling and forecasting
techniques available nowadays in a single environment. Once the initial barriers are overcome, a wide
panorama of modelling opportunities opens up.
18
2.3.8. Markov switching models14
Since the late 1980s, especially in macroeconomics and finance, the applications of dynamic econo-
metric modelling and forecasting techniques have increasingly relied on a special class of models that
accommodate regime shifts, Markov switching models (for short MSMs, also called hidden state Markov
models). The idea of MSMs is to relate the parameters of otherwise standard dynamic econometric frame-
works (such as systems of regressions, vector autoregressions, and vector error corrections) to one or more
unobserved state variables (say St , Vt ) that take K values and would capture the notion of systems go-
ing through phases or “regimes”. Such states follow a simple, discrete stochastic process, such as a M-th
order, recurrent, ergodic, non-reducible Markov chain, and are independent of the shocks of the model.
The typical assumption of M = 1 is without loss of generality, because higher order Markov chains can be
re-parameterised as a higher dimensional M = 1 chain by increasing K.
The mainstream literature (see, e.g., Hamilton 1990, or the textbook treatments by Kim et al. 1999, and
Guidolin and Pedio 2018) initially focused on time-homogeneous Markov chains (when the probabilities
of the state transitions are constant). However, the finance literature (Gray, 1996) has moved towards
time-heterogeneous MSMs, in which the transition matrix of the regimes may change over time, reacting
to lagged values of the endogenous variables, to lagged exogenous variables, or to the lagged values of the
state (in a self-exciting fashion).
For instance, in a K-regime MS VAR(p), all the parameters in the conditional mean (the vector of inter-
cepts and the p autoregressive matrices) may be assumed to take different, estimable values as a function
of St . Moreover, the covariance matrix of the system shocks may be assumed to depend on some state
variable, either the same as the mean parameters (St ) or an additional, specific one (Vt ), which may also
depend on lags of St . When a MS VAR model is extended to include exogenous regressors, we face a MS
VARX, of which MS regressions are just a special case.
Even though multivariate MSMs may suffer from issues of over-parameterisations that must be kept in
check, their power of fitting complex non-linearities is unquestioned because, as discussed by Marron and
Wand (1992), mixtures of normal distributions provide a flexible family that can be used to approximate
many distributions. Moreover, MSMs are known to capture key features of many time series (Timmer-
mann, 2000). For instance, differences in conditional means across regimes enter the higher moments such
as variance, skewness, and kurtosis; differences in means in addition to differences in variances across
regimes may generate persistence in levels and squared values of the series.
MSMs may be estimated by maximum likelihood, although other estimation methods cannot be ruled
out, like GMM (Lux, 2008). Typically, estimation and inference are based on the Expectation-Maximisation
algorithm proposed by Dempster et al. (1977), a filter that allows the iterative calculation of the one-
step ahead forecast of the state vector given the information set and a simplified construction of the
log-likelihood of the data. However, there is significant evidence of considerable advantages offered by
Bayesian approaches based on Monte Carlo Markov chain techniques to estimating multivariate MSMs
(see, for example, Hahn et al., 2010).
MSMs have been recently generalised in a number of directions, such as including regimes in condi-
tional variance functions (e.g., of a GARCH or DCC type, which poses significant statistical challenges in

14 This subsection was written by Massimo Guidolin.

19
writing the likelihood function; see Pelletier, 2006), and relating different sub-sets of parameters to state
vectors that follow a hierarchical structure, following Hamilton and Lin (1996).

2.3.9. Threshold models15

It is a well-known fact that financial and economic time series often display non-linear patterns. The
most common source of non-linearity is structural instability that may appear in the form of abrupt shifts
(breaks) or of recurrent regimes in the model parameters. In the latter case, such instability is stochastic, it
has a structure, and as such, it can be predicted. Accordingly, modelling economic and financial instability
has become an essential goal for econometricians since the 1970s.
One of the first and most popular models is the threshold autoregressive (TAR) model developed by
Tong (1978). A TAR model is an autoregressive model in which the parameters are driven by a state
variable St , which is itself a random variable taking K distinct values. In turn, the value assumed by St
depends on the value of the threshold variable qt resulting from a comparison with a set of K − 1 threshold
levels, qk∗ . The threshold variable qt can be exogenous or can be a lagged value of the dependent variable.
In the latter case, we speak of self-exciting threshold autoregressive (SETAR) models. Other choices of qt
include linear (see, for example, Chen et al., 2003; Chen and So, 2006; Gerlach et al., 2006) or non-linear
(see, for example, Chen, 1995; Wu and Chen, 2007) combinations of the lagged dependent variable or of
exogenous variables.
The TAR model has also been extended to account for different specifications of the conditional mean
function leading to the development of the threshold moving average (TMA: see, for example, Tong, 1978;
De Gooijer, 1998; Ling et al., 2007) and the threshold autoregressive moving average (TARMA: see, for
example, Ling, 1999; Amendola et al., 2006) models.
A criticism of TAR models is that they imply a conditional moment function that fails to be continuous.
To address this issue, Chan and Tong (1986) proposed the smooth transition autoregressive (STAR) model.
The main difference between TAR and STAR models is that, while a TAR imposes an abrupt shift from one
regime to the others at any time that the threshold variable crosses above (or below) a certain level, a STAR
model allows for gradual changes among regimes.
In its simplest form, a STAR is a two-regime model where the dependent variable yt is determined as the
weighted average of two autoregressive (AR) models. The weights are given by the transition function G,
which is a continuous function bounded between 0 and 1. The transition variable of G, xt , can be the lagged
endogenous variable yt−d for d ≥ 1 (see, for example, Teräsvirta, 1994), a (possibly non-linear) function of
it, or an exogenous variable. The transition variable can also be a linear time trend (xt = t), which generates
a model with smoothly changing parameters (see, for example, Lin and Teräsvirta, 1994). Popular choices
for G are the logistic function (which gives rise to the LSTAR model) and the exponential function (ESTAR),
although additional cumulative distribution functions are also possible. Notably, the simple STAR model
we have described can be generalised to have multiple regimes (see, for example, Van Dijk et al., 1999).
Threshold models are also applied to modelling and forecasting volatility; for instance, the GJR-GARCH
model of Glosten et al. (1993) can be interpreted as a special case of a threshold model. A few multivariate
extensions of threshold models also exist, such as vector autoregressive threshold models, threshold error

15 This subsection was written by Manuela Pedio.

20
correction models (see, for example, Balke and Fomby, 1997), and smooth transition vector error correction
models (see, for example, Granger and Swanson, 1996).

2.3.10. Peak over the theshold16

In the forecasting literature, Nikolopoulos (2020) argues that great attention has been given to mod-
elling fast-moving time series with or without cues of information available (Nikolopoulos et al., 2007).
Less attention has been given to intermittent/count series (see section 2.7), which are more difficult to
forecast given the presence of two sources of uncertainty: demand volume, and timing.
Historically there have been few forecasting methods developed specifically for such data (Syntetos
et al., 2015b). We believe that through a time series decomposition approach á la Leadbetter (1991) we can
isolate ‘peaks over threshold’ (POT) data points, and create new intermittent series from any time series
of interest. The derived series present almost identical characteristics with the series that Croston (1972)
analysed. In essence one could use such decomposition forecasting techniques to tackle much more diffi-
cult phenomena and problems coming from finance, politics, healthcare, humanitarian logistics, business,
economics, and social sciences.
Any time series can be decomposed into two sub-series: one containing the baseline (white swans)
and one containing the extreme values over an arbitrary-set or rule-based-set threshold (grey and black
swans) as proposed by Taleb (2008). Unfortunately, major decision-related risks and most of the underlying
uncertainty lay with these extremes. So, it is very important to be able to effectively model and forecast
them.
It is unlikely that any forecasting approach will accurately give the exact timing of the forthcoming
extreme event, but it will instead provide a satisfactory cumulative forecast over a long period of time. The
question still stands what can one do with this forecast? For earthquake data, although even if we know
that a major earthquake is going to hit a region, it is almost impossible to decide to evacuate cities, but still
we can influence and legislate the way structures are built and increase the awareness, training, prepared-
ness and readiness of the public; and also ensure enough capital on hold to cope with the aftermath of the
major event. For epidemics/pandemics there are clear implications, as we have evidenced with COVID-19,
on how proactively we can source and secure human resources, medical supplies, etc.
What is the current doctrine when forecasting in such a context: advanced probabilistic models. These
methods typically require a lot of data and reconstruct the distributions of the underlying phenomena.
These come with common successes and a plethora of constraints: big data sets needed for training the
models, high mathematical complexity, and invisibility to practitioners how these methods do actually
work and thus less acceptance in practice. Yet again, forecasting accuracy is the name of the game and thus
these forecasting methods are serious contenders for the task in hand.
Extreme Value Theory (EVT) analyses extreme deviations from statistical measures of central location
to estimate the probability of events that are more extreme than anything observed in the time series. This
is usually done in the following two ways (Nikolopoulos, 2020): (i) deriving maxima and/or minima series
as a first step and then having the Generalised Extreme Value Distribution fitted (often the number of
extreme events is limited), and (ii) isolating the values that exceed a threshold (point over threshold) that
can also lead to only a few instances extracted – so a very intermittent series in nature. The analysis involves

16 This subsection was written by Konstantinos Nikolopoulos.

21
fitting a Poisson distribution for the number of events in a basic time period and a second distribution –
usually a Generalised Pareto Distribution – for the size of the resulting POT values.

2.3.11. Forecasting with many variables17

Forecasts are often very simple models. The simplest are usually functions of deterministic terms,
or the forecast variable itself in previous time periods. Often these kinds of models can forecast better
than more complicated forecast models. By more complicated, we usually mean multivariate models –
regression models with multiple explanatory variables. Such models are often based on available theories
regarding the determination of the variable to be forecast, and are often referred to as structural models.
In a stationary world without structural change, then it would be anticipated that the best structural
model would provide the best forecasts, since it would provide the conditional mean of the data process
(see, for example, Clements and Hendry, 1998). In a non-stationary world of unit roots and structural
breaks, however, this need not be the case. In such situations, often simple forecast models can outperform
structural models, especially at short forecast horizons (see, for example, Hendry and Clements, 2001).
Multivariate forecast models require that explanatory variables also be forecast – or at least, scenarios
be set out for them. These may be simplistic scenarios, for example all explanatory variables take their
mean values. Such scenarios can play a useful role in formulating policy making since they illustrate in
some sense the outcomes of different policy choices.
Since the 1980s and Sims (1980), vector autoregressive (VAR) models have become ubiquitous in macroe-
conomics, and common in finance (see, for example, Hasbrouck, 1995). A VAR model is a set of linear
regression equations describing the evolution of a set of endogenous variables. Each equation casts each
variable as a function of lagged values of all the variables in the system.
Contemporaneous values of system variables are not included in VAR models for identification pur-
poses; some set of identifying restrictions are required, usually based on economic theory, and when im-
posed the resulting model is known as a structural VAR model.
VAR models introduce significantly greater levels of parameterisation of relationships, which increases
the level of estimation uncertainty. At the same time VAR models afford the forecaster a straightforward
way to generate forecasts of a range of variables, a problem when forecasting with many variables.
As with autoregressive methods, VAR models can capture a significant amount of variation in data
series that are autocorrelated, and hence VAR methods can be useful as baseline forecasting devices. VAR-
based forecasts are often used as a benchmark for complex models in macroeconomics like DSGE models
(see, for example, Del Negro and Schorfheide, 2006).
Indicator saturation is a method of detecting outliers and structural breaks by saturating a model with
different types of indicator, or deterministic variables (Johansen and Nielsen, 2009a; Castle et al., 2015a).
The flexibility of the approach is such that it has been applied in a wide variety of contexts, from volcanic
eruptions (Pretis et al., 2016) to prediction markets and social media trends (Vaughan Williams and Reade,
2016).
A particularly important and ever-expanding area of empirical analysis involves the use of panel data
sets. Outliers, structural breaks and split trends undoubtedly also exist in panel time series. The potential

17 This subsection was written by J. James Reade.

22
to test for common outliers and structural changes across cross sectional units would be useful, as would
the ability to allow individual units to vary individually, e.g., time-varying fixed effects.
Nymoen and Sparrman (2015) is the first application of indicator saturation methods in a panel context,
looking at equilibrium unemployment dynamics in a panel of OECD countries, but applications into the
panel context are somewhat constrained by computer software packages designed for indicator saturation.
The gets R package of Pretis et al. (2017, 2018) can be used with panel data.

2.3.12. Functional time series models18

Functional time series consist of random functions observed at regular time intervals. Functional time
series can be classified into two categories depending on if the continuum is also a time variable. On the
one hand, functional time series can arise from measurements obtained by separating an almost continuous
time record into consecutive intervals (e.g., days or years, see Horváth and Kokoszka, 2012). We refer to
such data structure as sliced functional time series, examples of which include daily precipitation data
(Gromenko et al., 2017). On the other hand, when the continuum is not a time variable, functional time
series can also arise when observations over a period are considered as finite-dimensional realisations of
an underlying continuous function (e.g., yearly age-specific mortality rates, see Li et al., 2020b).
Thanks to recent advances in computing storage, functional time series in the form of curves, images
or shapes is common. As a result, functional time series analysis has received increasing attention. For in-
stance, Bosq (2000) and Bosq and Blanke (2007) proposed the functional autoregressive of order 1 (FAR(1))
and derived one-step-ahead forecasts that are based on a regularised Yule-Walker equations. FAR(1) was
later extended to FAR(p), under which the order p can be determined via Kokoszka and Reimherr’s (2013)
hypothesis testing. Horváth et al. (2020) compared the forecasting performance between FAR(1), FAR(p),
and functional seasonal autoregressive models of Chen et al. (2019b).
To overcome the curse of dimensionality, a dimension reduction technique, such as functional principal
component analysis (FPCA), is often used. Aue et al. (2015) showed asymptotic equivalence between a FAR
and a vector autoregressive (VAR) model. Via an FPCA, Aue et al. (2015) proposed a forecasting method
based on the VAR forecasts of principal component scores. This approach can be viewed as an extension of
Hyndman and Shang (2009), in which principal component scores are forecast via a univariate time series
forecasting method. With the purpose of forecasting, Kargin and Onatski (2008) proposed to estimate
the FAR(1) model by using the method of predictive factors. Klepsch and Klüppelberg (2017) proposed
a functional moving average process and introduced an innovations algorithm to obtain the best linear
predictor. Klepsch et al. (2017) extended the VAR model to the vector autoregressive moving average
model and proposed the functional autoregressive moving average model.
Extending short-memory to long-memory functional time series analysis, Li et al. (2020a,b) considered
local Whittle and rescale-range estimators in a functional autoregressive fractionally integrated moving
average model. The models mentioned above require stationarity, which is often rejected. Horváth et al.
(2014) proposed a functional KPSS test for stationarity. Chang et al. (2016) studied nonstationarity of
the time series of state densities, while Beare et al. (2017) considered a cointegrated linear process in
Hilbert space. Nielsen et al. (2019) proposed a variance ratio-type test to determine the dimension of the

18 This subsection was written by Han Lin Shang.

23
nonstationary subspace in a functional time series. Li et al. (2020c) studied the estimation of the long-
memory parameter in a functional fractionally integrated time series, covering the functional unit root.
From a nonparametric perspective, Besse et al. (2000) proposed a functional kernel regression method
to model temporal dependence via a similarity measure characterised by semi-metric, bandwidth and ker-
nel function. Aneiros-Pérez and Vieu (2008) introduced a semi-functional partial linear model that com-
bines linear and nonlinear covariates. Apart from conditional mean estimation, Hörmann et al. (2013) con-
sidered a functional autoregressive conditional heteroscedasticity model for estimating conditional vari-
ance. Rice et al. (2020) proposed a conditional heteroscedasticity test for functional data. Kokoszka et al.
(2017) proposed a portmanteau test for testing autocorrelation under a functional generalised autoregres-
sive conditional heteroscedasticity model.

2.3.13. Bayesian forecasting with copulas19

Copulas provide intrinsic multivariate distribution structure that allows for modelling multivariate de-
pendence with marginal distributions as the input, making it possible to forecast dependent time series
and time series dependence. This review focuses on the Bayesian approach for copula forecasting. For
rigorous topics on copula introduction (Nelsen, 2006; Trivedi and Zimmer, 2007; Joe, 1997), copula mod-
elling techniques (Durante and Sempi, 2015), vine copulas (Joe, 2014), review on frequentist approaches
for copula-based forecasting (Patton, 2013), see the aforementioned references and the references therein.
The advantages of the Bayesian copula approach compared to the frequentist treatments are (i) the
Bayesian approach allows for jointly modelling the marginal models and the copula parameters, which
improves the forecasting efficiency (Joe, 2005; Li and Kang, 2018), (ii) probabilistic forecasting is naturally
implied with the Bayesian predictive density, and (iii) experts information can be seamlessly integrated
into forecasting via the priors’ setting.
Forecasting with copulas involves selecting between a different class of copulas. Common approaches
include Bayesian hypothesis testing where copula parameters are treated as nuisance variables (Huard
et al., 2006), or parsimonious modelling of covariance structures using Bayesian selection and model aver-
aging (Pitt et al., 2006; Smith, 2010; Min and Czado, 2011).
One particular interest in copulas is to forecast the dynamic dependencies between multivariate time
series. Time-varying copula construction is possible via (i) an autoregressive or GARCH form of de-
pendence parameters (Patton, 2006a; Lucas et al., 2014), (ii) factor copula construction (Oh and Patton,
2018; Tan et al., 2019) that simplifies the computation, (iii) the stochastic copula autoregressive model
(Almeida and Czado, 2012) that dependence is modelled by a real-valued latent variable, and (iv) covariate-
dependent copulas approach by parameterising the dependence as a function of possible time-dependent
covariates (Li and Kang, 2018) that also improves the forecasting interpretability. ARMA-like and GARCH-
like dependences in the tail can be considered as special cases of Li and Kang (2018).
In multivariate time series forecasting, unequal length of data is a common issue. One possible ap-
proach is to partition the copula parameters into elements relating only to the marginal distributions and
elements only relating to the copula (Patton, 2006b). For mixed frequency data, it is possible to decompose
the copula dependence structure into linear and nonlinear components. Then the high and low frequency
data is used to model the linear and nonlinear dependencies, respectively (Oh and Patton, 2016). Bayesian

19 This subsection was written by Feng Li.

24
data augmentation is also used to forecast multivariate time series with mixed discrete and continuous
margins (Smith and Khaled, 2012). For other treatments for discrete and continuous time series (see, for
example, Panagiotelis et al., 2012, 2017).
Bayesian approach for lower dimensional copula forecasting (d < 10) is straightforward with traditional
Gaussian copulas, Student’s-t copulas, Archimedean copulas, or pair copula combinations. In higher di-
mensional settings, special considerations are required to save the computational burden, such as low rank
approximation of covariates matrix (Salinas et al., 2019a) or factor copula models with stochastic loadings
(Creal and Tsay, 2015).
In the Bayesian setup, forecasting model performance is typically evaluated based on a K-fold out-of-
sample log predictive score (LPS: Geweke and Amisano, 2010), and out-of-sample Value-at-Risk (VaR) or
Expected Shortfall (ES) are particularly used in financial applications. The LPS is an overall forecasting
evaluation tool based on predictive densities, serving out-of-sample probabilistic forecasting. LPS is ideal
for decision makers (Geweke, 2001; Geweke and Amisano, 2010). The VaR gives the percentile of the
conditional distribution, and the corresponding ES is the expected value of response variable conditional
on it lying below its VaR.

2.3.14. Forecasting with DSGE models20

Dynamic Stochastic General Equilibrium (DSGE) models are the workhorse of modern macroeconomics
in both academia and policy-making institutions (for details, see: Paccagnini, 2017; Christiano et al., 2018).
For example, the European Central Bank uses the New Area-Wide Model introduced by Warne et al. (2010)
and the Federal Reserve Board has created the Estimated, Dynamic, Optimisation-based model (FRB/EDO)
as discussed in Chung et al. (2010). Developed as a response to Lucas (1976) critique of structural macroe-
conometrics models, DSGEs introduced microfundations to describe business cycle fluctuations. Initially
calibrated, estimated DSGEs have been employed in shocks identification and forecasting horseraces for
the last 15 years. Estimation became possible thanks to computational progress and adoption of Bayesian
techniques (for technical details, see: An and Schorfheide, 2007; Herbst and Schorfheide, 2016; Fernández-
Villaverde and Guerrón-Quintana, 2020). Bayesian estimation allows for attributing prior distributions,
instead of calibrating, and computing the posterior distribution for selected model parameters as well as
drawing from predictive density. The Smets and Wouters (2007) DSGE is the most popular framework
referred to in both research and policy literature. Proposed for the US economy, this medium-scale model
is a closed economy composed of households, labor unions, a productive sector, and a monetary policy
authority that sets the short-term interest rate according to a Taylor rule. These ingredients are mathe-
matically represented by a system of linear rational expectation equations. Using a solution algorithm (for
example, Blanchard and Kahn, 1980; Sims, 2002), researchers can write the model using the state-space
representation composed by the transition equation and the measurement equation. The latter matches the
observed data (in the Smets and Wouters: output growth rate, consumption, investment, wages, worked
hours, inflation, and short-term interest rate) with the model latent variables.
DSGE models forecasting performance is investigated along two dimensions: point forecast and density
forecast.

20 This subsection was written by Alessia Paccagnini.

25
 The point forecast is implemented by conducting both static and dynamic analysis, as described in Car-
dani et al. (2019). If the static analysis provides a unique forecast value, the dynamic analysis describes the
evolution of the prediction along the time dimension to investigate possible time-varying effects. Usually,
point predictions are compared using the Diebold and Mariano (1995) and the Clark and West (2006) tests
that compare predictions from two competing models The accuracy of the static analysis is based mainly
on Mean Absolute Error (MAE) and Root Mean Square Error (RMSE). MAE and RMSE are used to provide
a relative forecasting evaluation compared to other competitors. Following Clements and Hendry (1998),
Kolasa et al. (2012) apply the standard forecast unbiased test to assess if DSGEs are good forecasters in
the absolute sense. The accuracy of the dynamic analysis is based on the Fluctuation Test (for some DSGE
applications, see: Giacomini and Rossi, 2016; Cardani et al., 2019; Boneva et al., 2019). This test is based on
the calculation of RMSEs that are assessed to investigate if the forecasting performance can be influenced
by instabilities in the model parameters.
The density forecast is based on the uncertainty derived by the Bayesian estimation and it is com-
monly evaluated using the probability integral transform and the log predictive density scores (as main
references, Wolters, 2015; Kolasa and Rubaszek, 2015a). The statistical significance of these predictions is
evaluated using the Amisano and Giacomini (2007) test that compares log predictive density scores from
two competing models.

2.3.15. Low and high prices in volatility models21

Volatility models of financial instruments are largely based solely on closing prices; meanwhile, daily
low and high (LH) prices significantly increase the amount of information about the variation of returns
during a day. LH prices were used for the construction of highly efficient estimators of variance, so called
the range-based (RB) estimators (e.g., Parkinson, 1980; Garman and Klass, 1980; Rogers and Satchell, 1991;
Yang and Zhang, 2000; Fiszeder and Perczak, 2013). They, however, have a fundamental drawback, as they
neglect the temporal dependence of returns (like conditional heteroscedasticity) and do not allow for the
calculation of multi-period dynamic volatility forecasts.
In the last dozen or so years, numerous univariate dynamic volatility models have been constructed
based on LH prices. Some of them were presented in the review paper of Chou et al. (2015). These models
can be divided into four groups. The first one comprises simple models, used traditionally to describe
returns, but they are based on the price range or on the mentioned earlier RB variance estimators. They
include such models as random walk, moving average, exponentially weighted moving average (EWMA),
autoregressive (AR), autoregressive moving average (ARMA), and heterogeneous autoregressive (HAR).
The second group contains models which describe the conditional variance (or standard deviation) of re-
turns. It comprises models like GARCH-PARK-R (Mapa, 2003), GARCH-TR (Fiszeder, 2005), REGARCH
(Brandt and Jones, 2006), RGARCH (Molnár, 2016). The third group includes models which describe the
conditional mean of the price range. It means that in order to forecast variance of returns the results have
to be scaled. This group contains models like RB SV (Alizadeh et al., 2002), CARR (Chou, 2005), TARR
(Chen et al., 2008), CARGPR (Chan et al., 2012), STARR (Lin et al., 2012), and MSRB (Miao et al., 2013).
The last group is methodologically different because the estimation of model parameters is based on the

21 This subsection was written by Piotr Fiszeder.

26
sets of three prices, i.e., low, high and closing. This approach comprises the GARCH models (Lildholdt,
2002; Venter et al., 2005; Fiszeder and Perczak, 2016) and the SV model (Horst et al., 2012).
The development of multivariate models with LH prices has taken place in the last few years. They
can be divided into three groups. The first one includes models, used traditionally to describe returns
or prices, but they are based on the price range or RB variance estimators. They comprise such models
like multivariate EWMA, VAR, HVAR, and vector error correction (VEC). It is a simple approach, however
most models omit modelling the covariance of returns. The second group is formed by the multivariate
RB volatility models like RB-DCC (Chou et al., 2009), DSTCC-CARR (Chou and Cai, 2009), RR-HGADCC
(Asai, 2013), RB-MS-DCC (Su and Wu, 2014), DCC-RGARCH (Fiszeder et al., 2019), RB-copula (Chiang
and Wang, 2011; Wu and Liang, 2011). The third group includes the multivariate co-range volatility mod-
els like multivariate CARR (Fernandes et al., 2005), BEKK-HL (Fiszeder, 2018) and co-range DCC (Fiszeder
and Fałdziński, 2019). These models apply LH prices directly not only for the construction of variances of
returns but also for covariances.

2.3.16. Robust estimation and forecasting22

Estimators of time series processes can be dramatically affected by the presence of few aberrant obser-
vations which are called differently in the time series literature: outliers, spikes, jumps, extreme observa-
tions. If their presence is neglected, coefficients could be biasedly estimated. Biased estimates of ARIMA
processes will decrease the efficiency of predictions (Bianco et al., 2001). Moreover, as the optimal pre-
dictor of ARIMA models is a linear combination of observed units, the largest coefficients correspond to
observations near the forecast origin and the presence of outliers among these units can severely affect the
forecasts. The issue was first raised in the seminal paper by Fox (1972), who suggests a classification of
outliers in time series, separating additive outliers (AO) from innovation outliers (IO). The influence of dif-
ferent types of outliers on the prediction errors in conditional mean models (ARIMA models) is studied by
Chen and Liu (1993a,b) and Ledolter (1989, 1991), while the GARCH context is explored by Franses and
Ghijsels (1999) and Catalán and Trı́vez (2007). Abraham and George E. P. Box (1979) propose a Bayesian
model which reflects the presence of outliers in time series and allows to mitigate their effects on estimated
parameters and, consequently, improve the prediction ability. The main idea is to use a probabilistic frame-
work allowing for the presence of a small group of discrepant units.
Model specification for stationary and nonstationary processes is an issue to be considered very care-
fully when extreme observations are present. A procedure for the correct specification of models is intro-
duced by Tsay (1986) relying on iterative identification-detection-removal of cycles in the observed time
series contaminated by outliers. The same issue is tackled by Abraham and Chuang (1989); in this work
non-influential outliers are separated from influential outliers which are observations with high residuals
affecting parameter estimation. Tsay’s procedure has been later modified (Balke, 1993) to effectively detect
time series level shifts. The impulse- and step-indicator saturation approach is used by Marczak and Proi-
etti (2016) for detecting additive outliers and level shifts estimating structural models in the framework of
nonstationary seasonal series. They find that timely detection of level shifts located towards the end of the
series can improve the prediction accuracy.

22 This subsection was written by Luigi Grossi.

27
 All these works are important because outlier and influential observations detection are crucial for
improving the forecasting performance of models. Alternatively, the presence of extreme values can be
included in the process specification and, once parameters are estimated, the model can be used to obtain
bootstrap forecast intervals reflecting the possibility of large stochastic events in the future (Phillips, 1996).
The robust estimation of model parameters is another way to improve predictive accuracy without correct-
ing or removing outliers. Sakata and White (1998) introduce a new two-stage estimation strategy for the
conditional variance based on Hampel estimators and S-estimators. Park (2002) proposes a robust GARCH
model, called RGARCH exploiting the idea of least absolute deviation estimation. The robust approach is
also followed for conditional mean models by Gelper et al. (2009) who introduce a robust version of the ex-
ponential and Holt-Winters smoothing technique for prediction purposes and by Cheng and Yang (2015)
who propose an outlier resistant algorithm developed starting from a new synthetic loss function. Very
recently, Beyaztas and Shang (2019) have introduced a robust forecasting procedure based on weighted
likelihood estimators to improve point and interval forecasts in functional time series contaminated by the
presence of outliers.

2.3.17. Robust equilibrium-correction forecasting devices23

The use of equilibrium-correction models is ubiquitous in forecasting. Hendry (2010) notes that in
economics this class commonly includes models with explicit equilibrium-correction mechanisms such
as vector equilibrium-correction models (VEqCM) as well as models with implicit equilibrium-correction
(or long-run mean reversion) mechanisms such as vector auto-regression (VAR) models, dynamic factor
models (DFM), dynamic stochastic general-equilibrium (DSGE) models, most models of the variance, and
almost all regression equations. This class of forecast model is also prevalent across many other disciplines.
For example, Pretis (2020) illustrates that there is an equivalence between physical energy balance models,
which are used to explain and predict the evolution of climate systems, and VEqCMs.
Despite their wide-spread use in economic modelling and forecasting, equilibrium-correction models
often produce forecasts that exhibit large and systematic forecast errors. Clements and Hendry (1998,
1999) showed that forecasts from equilibrium-correction models are not robust to abrupt changes in the
equilibrium. These types of regime changes are very common in macroeconomic time series (Hamilton,
2016) and can cause the forecasts from many models to go off track. Therefore, if for example, there is a
change in the equilibrium towards the end of the estimation sample, forecasts from this class of models
will continue to converge back to the previous equilibrium.
In general, the forecasts from equilibrium-correction models can be robustified by estimating all model
parameters over smaller or more flexible sub-samples. Several studies have proposed general procedures
that allow for time-variation in the parameters; see, for example, Pesaran et al. (2013), Giraitis et al. (2013),
and Inoue et al. (2017). This allows for an optimal or more adaptive selection of model estimation windows
in order generate forecasts after structural breaks have occured.
An alternative approach for robustifying forecasts from equilibrium-correction models is to focus on
the formulation and estimation of the equilibrium. Hendry (2006) shows that differencing the equilibrium-
correction mechanism can improve the forecasts by removing aspects which are susceptible to shifts. How-
ever, differencing the equilibrium also induces additional forecast-error variance. Castle et al. (2010) show

23 This subsection was written by Andrew B. Martinez.

28
that it is beneficial to update the equilibrium or to incorporate the underlying structural break process. Al-
ternatively, Castle et al. (2015b) show that there can be large forecast gains from smoothing over estimates
of the transformed equilibrium. Building on this, Martinez et al. (2019) show that there are many pos-
sible transformations of the equilibrium that can improve the forecasts. Several of these transformations
imply that the equilibrium-correction model collapses to different naive forecast devices whose forecasts
are often difficult to beat. By replacing the equilibrium with smooth estimates of these transformations,
it is possible to outperform the naive forecasts at both short and long forecast horizons while retaining
the underlying economic theory embedded within the equilibrium-correction model. Thus, it is possible
to dramatically improve forecasts from equilibrium-correction models using targeted transformations of
the estimated equilibrium so that it is less susceptible to the shifts which are so damaging to the model
forecasts.

2.3.18. Forecasting with data subject to revision24

When a forecast is made today of the future value of a variable, the forecast is necessarily ‘real time’
– only information available at the time the forecast is made can be used. The forecasting ability of a
model can be evaluated by mimicking this setup – generating forecasts over some past period (so outcomes
known) only using data known at each forecast origin. As noted by Clements and Hendry (2005), out-of-
sample forecast performance is the gold standard. Sometimes the analysis is pseudo real time. At a given
forecast origin t, forecasts are constructed only using data up to the period t, but the data are taken from
the latest-available vintage at the time the study is undertaken. The improved availability of real-time
databases has facilitated proper real-time studies.25 At time t the data are taken from the vintage available
at time t. Data revisions are often important, and occur because statistical agencies strive to provide timely
estimates which are based on incomplete source data (see, for example, Fixler and Grimm, 2005, 2008;
Zwijnenburg, 2015).
There are a number of possible real-time approaches. The conventional approach is to estimate the
forecasting model using the latest vintage of data available at time t. Suppose the vintage-t contains data
t t t
for time periods up to t − 1, denoted . . . , yt−3 , yt−2 , yt−1 . The observation for time t − 1 is a first estimate, for
t − 2 a second estimate, and so on, such that data for earlier periods will have been revised many times.
Hence the model will be estimated on data of different maturities, much of which will have been revised
many times. But the forecast will typically be generated by feeding into the model ‘lightly-revised’ data for
the most recent time periods. The accuracy of the resulting forecasts can be improved upon (in principle)
by taking into account data revisions (see, for example, Koenig et al., 2003; Kishor and Koenig, 2012;
Clements and Galvão, 2013b).
Koenig et al. (2003) suggest using real-time-vintage (RTV) data to estimate the model. The idea is to
use early estimates of the data to estimate the model, so that the model is estimated on ‘lightly-revised’
data that matches the maturity of the forecast-origin data that the forecast is conditioned on.
Other approaches seek to model the data revisions process along with the fully-revised true values of
the data, as in Kishor and Koenig (2012), Cunningham et al. (2009), and Jacobs and van Norden (2011).

24 This subsection was written by Michael P. Clements.

25 For example, the Federal Reserve Bank of Philadelphia maintain a real-time data set covering a number of US macro-variables,
at: https://www.philadelphiafed.org/research-and-data/real-time-center/real-time-data/, and see Croushore and Stark (2001).

29
Reduced form models that avoid the necessity of estimating unobserved components have adapted the
vector autoregression (VAR) of Sims (1980) to jointly model different observed vintages of data. Following
Patterson (1995), Garratt et al. (2008) work in terms of the level of the log of the variable, Ytt+1 , and
 0
model the vector given by Zt+1 = Ytt+1 − Yt−1
t t+1
, Yt−1 t
− Yt−1 t+1
, Yt−2 t
− Yt−2 . Clements and Galvão (2012, 2013a)
and Carriero et al. (2015) minimise the effects of benchmark revisions and re-basing by modelling ‘same-
 0
vintage-growth rates’, namely Zt+1 = ytt+1 , yt−1
t+1 t+1
, . . . , yt−q+1 , where ytt+1 = Ytt+1 − Yt−1
t+1
, and q denotes the
greatest data maturity.
Galvão (2017) shows how forecasts of fully-revised data can be generated for Dynamic stochastic gen-
eral equilibrium (DSGE) models (for example, Del Negro and Schorfheide, 2013), by applying the approach
of Kishor and Koenig (2012). Clements (2017) argues that improvements in forecast accuracy might be ex-
pected to be greater for interval or density forecasts than point forecasts, and this is further explored by
Clements and Galvão (2017).
Surveys on data revisions and real-time analysis, including forecasting, are provided by Croushore
(2006, 2011b,a) and Clements and Galvão (2019).

2.3.19. Innovation diffusion models26

Forecasting the diffusion of innovations is a broad field of research, and influential reviews on the topic
have highlighted its importance in many disciplines for strategic or anticipative reasons (Mahajan et al.,
1990; Meade and Islam, 2006; Peres et al., 2010). Large-scale and fast diffusion processes of different
nature, ranging from the spread of new epidemics to the adoption of new technologies and products,
from the fast diffusion of news to the wide acceptance of new trends and social norms, are demanding
a strong effort in terms of forecasting and control, in order to manage their impact into socio-economic,
technological and ecological systems.
The formal representation of diffusion processes is often based on epidemic models, under the hy-
pothesis that an innovation spreads in a social system through communication among people just like an
epidemics does through contagion. The simplest example is represented by the (cumulative) logistic equa-
tion that describes a pure epidemic process in a homogeneously mixing population (Verhulst, 1838). The
most famous and employed evolution of the logistic equation is the Bass model (Bass, 1969) developed
in the field of quantitative marketing and soon become a major reference, due to its simple and power-
ful structure. The Bass model (BM) describes the life-cycle of an innovation, depicting its characterising
phases of launch, growth/maturity, and decline, as result of the purchase decisions of a given cohort of
potential adopters. Mathematically, the model is represented by a first order differential equation, describ-
ing a diffusion process by means of three parameters: the maximum market potential, m, assumed to be
constant along the whole diffusion process, and parameters p and q, referring respectively to two distinct
categories of consumers, the innovators, identified with parameter p, adopting for first, and the imitators,
adopting at a later stage by imitating others’ behaviour and thus responsible for word-of-mouth dynamics.
In strategic terms, crucial forecasts are referred to the point of maximum growth of the life cycle, the peak,
and the point of market saturation.
Innovation diffusion models may be also used for post-hoc explanations, helping understand the evolu-
tion of a specific market and its response to different external factors. Indeed, one of the most appealing

26 This subsection was written by Mariangela Guidolin.

30
characteristics of this class of models is the possibility to give a simple and nice interpretation to all the
parameters involved. In this perspective, a valuable generalisation of the BM was proposed in Bass et al.
(1994) with the Generalised Bass Model (GBM). The GBM enlarges the BM by multiplying its hazard rate
by a very general intervention function x(t), assumed to be non-negative, which may account for exogenous
shocks able to change the temporal dynamics of the diffusion process, like marketing strategies, incentive
mechanisms, change in prices and policy measures.
Another generalisation of the BM and the GB¡, relaxing the assumption of a constant market potential
was proposed in Guseo and Guidolin (2009) with the GGM. This model postulates a time-dependent mar-
ket potential, m(t), which is function of the spread of knowledge about the innovation, and thus assumes
that a diffusion process is characterised by two separate phases, information and adoption. The GGM
allows a significant improvement in forecasting over the simpler BM, especially through a more efficient
description of the first part of the time series, often characterised by a slowdown pattern, as noticed by
Guseo and Guidolin (2011).

2.3.20. The natural law of growth in competition27

As early as in 1925 Alfred J. Lotka demonstrated that manmade products diffuse in society along S-
shaped patterns similar to those of the populations of biological organisms (Lotka, 1925). Since then
S curve logistic descriptions have made their appearance in a wide range of applications from biology,
epidemiology and ecology to industry, competitive substitutions, art, personal achievement and others
(Fisher and Pry, 1971; Marchetti, 1983; Meade, 1984; Modis, 1992). In fact, logistic growth can be detected
whenever there is growth in competition, and competition can be generalised to a high level of abstraction,
e.g., diseases competing for victims and all possible accidents competing for the chance to be materialised.
S curves enter as modular components in many intricate natural patterns. One may find S curves inside
S curves because logistics portray a fractal aspect; a large S curve can be decomposed in a cascade of
smaller ones (Modis, 1994). One may also find chaos by rendering the logistic equation discrete (Modis
and Debecker, 1992). Finally, logistics sit in the heart of the Volterra-Lotka equations, which describe the
predator-prey relations and other forms of competition. In its full generality, the logistic equation, in a
discrete form, with cross terms to account for all interrelations between competing species, would give
a complete picture in which growth in competition, chaos, self-organisation, complex adaptive systems,
autopoiesis, and other such academic formulations, all ensue as special cases (Modis, 1997).
Each S curve has its own life cycle, undergoing good and bad “seasons” (see figure 1). A large set of
behaviours have been tabulated, each one best suited for a particular season (Modis, 1998). Becoming
conservative – seeking no change – is appropriate in the summer when things work well. But excellence
drops in second place during the difficult times of winter – characterised by chaotic fluctuations – when
fundamental change must take place. Learning and investing are appropriate for spring, but teaching,
tightening the belt, and sowing the seeds for the next season’s crop belong in the fall.
S curves can also serve simply qualitatively to obtain rare insights and intuitive understanding. The
seasons metaphor dictates that in spring the focus is on what to do, whereas in fall the emphasis shifts
to the how. For example, the evolution of classical music can be visualised as a large-timeframe S curve
beginning sometime in the fifteenth century and reaching a ceiling in the twentieth century (see figure

27 This subsection was written by Theodore Modis.

31
 Work ethic Vertical Tighten belt
Fundamental integration
change Learning Benchmarking
Conservatism
Segmentation Product
innovation Leadership Back to
Decentralization basics
Build Excellence
Entrepreneurship capacity Process
TQM innovation
Business Continuous
process improvement Face lift
reengineering Specialists
evolve into
Fire bureaucrats Hire bureaucrats Hire
specialists generalists

Winter Spring Summer Fall Winter

Figure 1: Typical attributes of a growth cycle’s “seasons”. Adopted from Modis (1998) with the permission of the author.

2). In Bach’s time composers were concerned with what to say. The value of their music is in its archi-
tecture and as a consequence it can be interpreted by any instrument, even by simple whistling. But two
hundred years later composers such as Debussy wrote music that depends crucially on the interpretation,
the how. Classical music was still “young” in Bach’s time but was getting “old” by Debussy’s time. No
wonder Chopin is more popular than Bartók. Chopin composed during the “summer” of music’s S curve
when public preoccupation with music grew fastest. Around that time composers were rewarded more
handsomely than today. The innovations they made in music were assimilated by the public within a short
period of time because the curve rose steeply and would rapidly catch up with each innovation. But today
the curve has flattened and composers are given limited space. If they make even small innovations they
find themselves above the curve and there would not be any time in the future when the public will ap-
preciate their work. On the other hand, if they do not innovate, they will not be saying anything new. In
either case today’s composers will not be credited with an achievement. An S curve constructed using only
qualitative arguments seems accurate and informative.

2.3.21. Synchronic and diachronic competition28

Competition in a social system refers to the presence of antagonists that contend a resource. Separate
time series describe rate variables of the individually acquired resource representing local oscillatory be-
haviour with nonlinear trends, unexpected deviations and saturating effects. The use of standard time se-
ries analysis may be misleading for crucial omissions. A nonlinear trend cannot be ‘absorbed’ by low-order
differences. Consequently, seasonal and autoregressive components cannot be correctly identified. The
analytic representation of competition follows different approaches. Systems analysis exploits differential
equations describing rate variables as functions of all variables and part of their interactions. Complex
systems analysis (Boccara, 2004) refers to a class of agents that interact through local transition rules pro-
ducing competition as an emergent behaviour. Through mean field approximations, previous approach

28 This subsection was written by Renato Guseo.

32
Figure 2: The evolution of classical music. The vertical axis could be something like “importance”, “public appreciation”, or “public
preoccupation with music”, (cumulative). Adopted from Modis (2013) with the permission of the author.

may be frequently reduced to a system of differential equations. Competition is synchronic if antagonists

reach a common environment simultaneously; diachronic when the late entrant is introduced later. First
historical contributions are due to Lotka (1920) and Volterra (1926), which independently obtained a class
of predator-prey synchronic models.
A generalised version of the Lotka-Volterra model, LV, is provided by Abramson and Zanette (1998).
In a duopoly (Morris and Pratt, 2003) the equations explicit the role of local carrying capacities or market
potentials, the inhibiting strength of the rivals in accessing the residual resource, and the rate modulation
of interaction or word-of-mouth (WOM) effects between the specific cumulative resource and its untrans-
formed residual. The residual resource is essentially specific and interactions exclude first and second
order components of the rivals. Usually, Lotka-Volterra models do not have closed form solutions. In both
cases, staked form of equations or solutions allows a first-stage nonparametric inference based on nonlin-
ear least squares (NLS) avoiding strong assumptions on the stochastic distributions of error components.
Short-term refining may be grounded on a Seasonal Autoregressive Moving Average with exogenous in-
put (SARMAX) representation. Outstanding forecasts are obtained including the first-stage solution as
‘exogenous’ input.
A different synchronic approach, Givon-Bass (GB) that extends diffusion of innovations theory (Bass,
1969; Givon et al., 1995) is given in Bonaldo (1991) through equations that involve parametric compo-
nents of global interaction. The residual resource is completely accessible to all competitors, and the rate
equations introduce distributed seeding effects. The GB model has a closed form solution that was inde-
pendently published by Krishnan et al. (2000). The model and related advances by Savin and Terwiesch
(2005) and Guseo and Mortarino (2010) were generalised to the diachronic case in Guseo and Mortarino
(2012), defining a competition and regime change diachronic model (CRCD). The carrying capacity has
two fixed regimes before and during competition. Similar aspects concern seeding and interaction effects.
CRCD has a closed form solution. A relevant extension of CRCD is discussed in Guseo and Mortarino
(2014) introducing within-brand and cross-brand WOM effects not included in Lotka-Volterra approach.
The unrestricted unbalanced diachronic competition model (UCRCD) is defined with these new compet-

33
ing effects. It does not have a closed form solution for general ‘discriminations’ parameters. The restricted
UCRCD case, for restricted discriminations, has two different solutions. The model assumes a local invari-
ance of carrying capacity. In Guseo and Mortarino (2015) this assumption was weakened, introducing a
dynamic carrying capacity approach (Guseo and Guidolin, 2009, 2011). The CRCD is a special submodel
of UCRCD.

2.3.22. Estimation and representation of uncertainty29

Forecasting uncertainty consists in estimating the possible range of forecast errors (or true values) in
the future and the most widely adopted representation is a forecast interval (Patel, 1989). The forecast
interval indicates a range of values and the respective probability, which is likely to contain the true value
(which is yet to be observed) of the response variable. Since for a specific lead-time the forecast inter-
val only encompasses information from a marginal distribution, it can also be named marginal forecast
interval (MFI). A MFI can be obtained from: parametric distribution, such as a Gaussian distribution
with conditional mean and variance estimated with a Generalised ARCH model (Baillie and Bollerslev,
1992); non-parametric distribution, e.g., obtained with conditional kernel density estimation (Hyndman
et al., 1996); directly estimated with statistical learning methods, such as quantile regression (Taylor and
Bunn, 1999) or bootstrapping (Masarotto, 1990), or with machine learning algorithms like quantile ran-
dom forests (Meinshausen, 2006).
For multi-step ahead forecasting problems, in particular when information about forecast uncertainty
is integrated in multi-period stochastic optimisation (Dantzig and Infanger, 1993), information about the
temporal dependency structure of forecast errors (or uncertainty) is a fundamental requirement. In this
case, the concept of simultaneous forecast intervals (SFI) can be found in the statistical literature (Chew,
1968). SFI differ from MFI since take into account the temporal interdependency of forecast errors and
are constructed to have the observed temporal trajectory of the response variable fully contained inside
the forecast intervals during all lead-times of the time horizon. The number of works that cover SFI is
lower when compared to the MFI, but some examples are: methods based on Bonferroni- and product-type
inequalities applied time series forecasting models likes ARIMA and Holt-Winters (Ravishanker et al.,
1991); combination of bootstrap replications and an heuristic optimisation procedure to find an envelope
of the temporal trajectories closest to the deterministic forecast (Staszewska-Bystrova, 2011); sampling
forecast errors at different horizons and estimate the SFI with the empirical Mahalanobis distance (Jordá
et al., 2013).
Advances in Operations Research for decision-making problems under uncertainty imposed new re-
quirements in forecast uncertainty estimation and representation. On one hand, stochastic optimisation
requires a scenario representation for forecast uncertainty (Powell, 2019). This motivated research in meth-
ods that generate uncertainty forecasts represented by random vectors (term used in statistics) or path
forecasts (term used in econometrics), such as parametric copula combined with MFI (Pinson et al., 2009),
parametric dynamic stochastic model (Li and Chan, 2011) or epi-spline basis functions (Rios et al., 2015).
On the other hand, robust optimisation does not make specific assumptions on probability distributions
and the uncertain parameters are assumed to belong to a deterministic uncertainty set. Hence, some au-
thors proposed new methods to shape forecast uncertainty as polyhedral or ellipsoidal regions to enable

29 This subsection was written by Ricardo Bessa.

34
a direct integration of forecasts in this type of optimisation problem (Bertsimas and Pachamanova, 2008;
Golestaneh et al., 2019).
Finally, communication of forecast uncertainty (e.g., MFI, SFI, random vectors) to decision-makers re-
quires further attention since it remains as a major bottleneck for a wide adoption by industry, particularly
in uses cases with multivariate time series (Akram et al., 2015) and adverse weather events (Ramos et al.,
2010). Please also see section 3.7.8.

2.3.23. Forecasting under fat tails30

A non-negative continuous random variable X is fat-tailed, if its survival function S(x) = P (X ≥ x)
is regularly varying, that is to say if S(x) = L(x)x−α , where L(x) is a slowly varying function, for which
L(tx)
limx→∞ L(x)
= 1 for t > 0 (Embrechts et al., 2013). The parameter α is known as the tail parameter, and it
governs the thickness of the tail – the smaller α the fatter the tail – and the existence of moments, so that
E[X p ] < ∞ if and only if α > p. Often α is re-parametrised as ξ = 1/α.
Fat tails are omnipresent in nature, from earthquakes to floods, and they are particularly common
in human-related phenomena like financial markets, insurance, pandemics, and wars (see, for example,
Mandelbrot, 1983; Taleb, 2020, and references therein).
Forecasting fat-tailed random variables is therefore pivotal in many fields of life and science. However,
while a basic coverage of the topic is available in most time series and risk management manuals (e.g.,
Shumway and Stoffer, 2017; McNeil et al., 2015), the profound implications of fat tails are rarely taken
into consideration, and this can generate substantial errors in forecasts.
As observed in Taleb et al. (2020), any statistical forecasting activity about the mean – or another quan-
tity – of a phenomenon needs the law of large numbers (LLN), which guarantees the convergence of the
sample mean at a given known rate, when the number of observations n grows.
Fat-tailed phenomena with tail parameter α ≤ 1 are trivially not predictable. Since their theoretical
mean is not defined, the LLN does not work, for there is nothing the sample mean can converge to. This also
applies to apparently infinite-mean phenomena, like pandemics and wars, i.e., extremely tricky objects of
study, as discussed in Cirillo and Taleb (2016a). In similar situations, one can rely on extreme value theory
to understand tail risk properties, but should refrain from standard forecasting.
For random variables with 1 < α ≤ 2, the LLN can be extremely slow, and an often unavailable number
of observations is needed to produce reliable forecasts. Even for a well-behaved and non-erratic phe-
nomenon, we all agree that a claim about the fitness or non-fitness of a forecasting approach, just on the
basis of one single observation (n = 1), would be considered unscientific. The fact is that with fat-tailed
variables that “n = 1” problem can be made with n = 106 observations (Embrechts et al., 2013; Taleb,
2020). In the case of events like operational losses, even a larger n → ∞ can still be just anecdotal (Cirillo
and Taleb, 2016a).
According to Taleb (2020), owing to preasymptotics, a conservative heuristic is to manage variables
with α ≤ 2.5 as practically unpredictable. Their sample average is indeed too unstable and needs too many
observations for forecasts to be reliable in a reasonable period of time. For α > 2.5, conversely, forecasting
can take place, and the higher α the better. In any case, more research is strongly needed and desirable.

30 This subsection was written by Pasquale Cirillo.

35
 Observe that even discussing the optimality of any alarm system (see, for example, Turkman and Turk-
man, 1990; Svensson et al., 1996) based on average forecasts would prove meaningless under extremely fat
tails (α ≤ 2), when the LLN works very slowly or does not work. In fact, even when the expected value is
well-defined (i.e., 1 < α < 2), the non-existence of the variance would affect all the relevant quantities for
the verification of optimality (De Mare, 1980), like for instance the chance of undetected events. For all
these quantities, the simple sample estimates commonly used would indeed be misleading.

2.4. Variable and model selection

2.4.1. Leading indicators and Granger causality31
Leading (economic) indicators are variables that try to capture changes in the development of economic
activity before those changes materialise. Typically, market participants, policy makers, the general public,
etc. are interested in which direction the economy of a country is developing ahead of the publication
date of the respective quarterly GDP figures, for which these leading indicators play a crucial role. This
is of particular relevance for short-term forecasting and nowcasting of such a (low-frequency) variable
characterised by a publication lag.
For some leading indicators, single variables are combined into a composite index such as The Confer-
ence Board Leading Economic Index published by The Conference Board for the United States. It consists
of averages of ten variables, including average weekly hours in manufacturing, manufacturers’ new or-
ders, private housing building permits, and average consumer expectations (The Conference Board, 2020).
Other leading indicators consist of one variable only and are purely survey-based, such as the monthly ifo
Business Climate Index published by the ifo Institute for Germany, which covers the appraisal of current
and future expected business of executives of approximately 9,000 German firms across industries (ifo
Institute, 2020).
If a specific leading indicator is able to capture certain economic changes before they happen, such
a leading indicator is said to Granger-cause (or predictively cause) a forecast variable (Granger, 1969),
implying that a “cause” cannot happen after an effect. In (theoretical) econometric terms and following
the description of Lütkepohl (2005), a forecast variable is Granger-caused by a leading indicator if for at
least one forecast horizon h the following inequality holds: the mean square error of the optimal h-step-
ahead predictor of the forecast variable, which includes the information contained in the leading indicator
in the set of all relevant information to the forecaster available at the forecast origin, must be smaller
than the mean square error of that h-step-ahead predictor of the forecast variable without the information
contained in said leading indicator (Lütkepohl, 2005).
Nonetheless, the concept of Granger causality also has its limitations. The notion that a “cause” cannot
happen after an effect is only a necessary, yet not a sufficient condition for causality, hence the reason why
“cause” has been written with quotation marks in this section. If one ignored this fact and equated causality
with Granger causality, they would commit an informal logical fallacy called Post hoc ergo propter hoc (i.e.,
after this, therefore because of this; Walton et al., 2008). A bold example of this fallacy would be, “Because
the birds migrate, winter is coming”. This is a fallacy, as winter would come at about the same time every
year, no matter if there were migrating birds or not. Moreover, hardly any economic activity is monocausal.

31 This subsection was written by Ulrich Gunter.

36
 There are also different types of formal statistical Granger causality tests available for different data
structures that are implemented in typical statistics/econometrics software packages. For the simple case
of two variables (Granger, 1969), say the forecast variable and the leading indicator, the null hypothesis of
a bivariate Granger causality test is that the leading indicator does not Granger-cause the forecast variable.
Under this null hypothesis, the F-test statistic on the joint impact of the coefficients of the past realisations
of the leading indicator employed as explanatory variables in a multiple linear regression with the forecast
variable as dependent variable and its past realisations as additional explanatory variables will not be
statistically significantly different from zero. The maximum lag for past realisations would be optimally
determined, for instance, by some information criterion (e.g., AIC, BIC).

2.4.2. Model complexity32

A simple model must be easily understood by decision-makers. On the contrary, relationships in a
complex model are opaque for its users. In this context, complexity is not measured solely by the number
of parameters, but also by the functional form of the relationships among variables.
Complex models are commonly believed to deliver better forecasts, as they well describe sophisticated
economic structures and offer good fit the data. Consequently, they are favoured by researchers, decision-
makers and academic journals. However, empirical studies provide little evidence about their forecasting
superiority over simple models. This can be explained using the bias-variance trade-off framework, in
which the mean squared error can be decomposed into

MSE = noise + variance + bias2 .

Noise is driven by the random term in the model. It cannot be avoided, even if we know the true DGP.
Variance is caused by the need to estimate the model parameters, hence its value increases with model
complexity and declines with the sample size. Bias is predominantly related to model mis-specification,
which is most likely to occur for simple methods. The implications of the above framework are twofold: (i)
the relationship between model complexity and MSE is U-shaped and (ii) the optimal model complexity
increases with the sample size.
The illustration of the bias-variance trade-off for a simple autoregressive model is provided by Ca’ Zorzi
et al. (2016). For a stationary DGP, the accuracy of forecasts from the random walk and a calibrated
AR(1) model tend to be higher than that from an estimated AR(1). This result explains why in numerous
studies the random walk is a tough benchmark as well as why a simple, calibrated AR(1) model can be
successful in forecasting inflation (Faust and Wright, 2013), exchange rates (Ca’ Zorzi et al., 2017) or oil
prices (Rubaszek, 2020) compared to a number of complex competitors.
Wide support to the view that model simplicity improves forecasting performance is presented by
Green and Armstrong (2015), in an introductory article to the special issue of Journal of Business Research
“Simple versus complex forecasting”, as well as the results of M1 and M2 competitions (Makridakis et al.,
1982, 1993). Stock and Watson (1998) also show that for most US monthly series complex non-linear
autoregressive models deliver less accurate forecasts than their linear counterparts. On the contrary, the

32 This subsection was written by Michał Rubaszek.

37
results of the M3 and M4 competitions tend to favour more complex models (Makridakis and Hibon, 2000;
Makridakis et al., 2020c).
Why are then complex models preferred to simple ones, if the latter deliver more accurate forecasts?
Brighton and Gigerenzer (2015) claim that there is a tendency to overemphasize the bias component and
downplay the role of variance. This behaviour is leading to an implicit preference towards more complex
forecasting methods, which is called by the authors as “bias bias”. To avoid it, one can follow the golden
rules of forecasting, which says: be conservative in the choice of over-ambitious models and be wary of the
difficulty to forecast with complex methods (Armstrong et al., 2015).

2.4.3. Variable selection33

References to ‘big data’ have become somewhat ubiquitous both in the media and in academic literature
in recent years. Whilst in some disciplines (for example Finance) it has become possible to observe time
series data at ever higher frequency, it is in the cross section that the amount of data available to analysts
has seen exponential growth.
Ordinary Least Squares (OLS) is the standard tool for data analysis and prediction, but is well known
to perform poorly when there many potential explanatory variables; Bishop (2006) sets out clearly why
this is so. In situations where there is no obvious underlying model to suggest which of a potentially large
set of candidate variables to focus on, the researcher needs to add new tools to the tool kit.
There are two principal sets of approaches to this problem. The first seeks to reduce the model dimen-
sions by summarising the predictors in to a much smaller number of aggregated indices or factors. The
second is to employ a regularisation strategy to reduce the effective dimension of the problem by shrinking
the size of the regression coefficients. Some strategies reduce a subset of these coefficients to zero, removing
some variables from the model entirely and these can hence be described as variable selection procedures.
Such procedures may be applicable either when a problem is believed to be truly sparse, with only
a small number of variables having an effect, or alternatively when a sparse model can provide effective
forecasts, even though the underling system is in fact more complex.
In the frequentist framework, the Least Absolute Shrinkage and Selection Operator (LASSO) procedure
of Tibshirani (1996) has proven effective in many applications. The LASSO requires choice of an additional
regularisation parameter (usually selected by some statistical criteria or via a cross validation process).
Various refinements to the original LASSO procedure have been developed, see in particular Rapach et al.
(2019) for a recent forecasting application. An alternative frequentist approach to variable selection and
shrinkage is the Complete Subset Regression (CSR) model of Elliott (2015), where separate OLS regressions
are run on all possible combinations of potential regressors, with forecasts generated by averaging across
the entire set of models. Kotchoni et al. (2019) combine CSR with LASSO in a comprehensive empirical
economic forecasting exercise.
Where the underlying process is not properly sparse (see Giannone and Primiceri, 2017, for a discus-
sion), it is perhaps more natural to work in a Bayesian framework where samples can be drawn from the
variable selection part of model reflecting the estimated probability of inclusion of each variable. The ap-
propriate degree of regularisation can also be selected in a similar way. Forecasts are then constructed as
weighted averages over several sparse models. This approach has proven to be very effective in practice,

33 This subsection was written by Ross Hollyman.

38
for example in fitting Bayesian Vector Auto Regressions to forecast economic time series. Early examples
include George and McCulloch (1993) and Mitchell and Beauchamp (1988), which use binary indicators
to select variables in to the model. More recent approaches use continuous random variables to achieve a
similar effect, making computation more tractable. Examples include the Horeshoe Prior (Carvalho et al.,
2010; Piironen and Vehtari, 2017) and the LN-CASS prior of Thomson et al. (2019). Cross (2020) is a recent
example of an economic forecasting exercise using several such models.

2.4.4. Model selection34

Even when selecting for forecasting facing multiple breaks at unknown times, the natural target for
model selection is the data generating process (DGP: Hendry, 1995). However, taxonomies of all possible
sources of forecast errors from estimated models that are not the DGP show only a couple of mistakes
determine forecast failures (systematic departures between forecasts and later outcomes), namely mis-
measured forecast origins and unanticipated location shifts (Clements and Hendry, 1999; Hendry and
Mizon, 2012). Failure to capture these can distort estimation and forecasts. When facing an unknown
number of in-sample contaminating outliers and multiple breaks at unknown times, selection must be
jointly over both observations and variables, requiring computer learning methods (Hendry and Doornik,
2014; Castle et al., 2020b). Autometrics, a multiple-path block-search algorithm (see Doornik, 2018), uses
impulse (IIS: Hendry et al., 2008 and Johansen and Nielsen, 2009b) and step (SIS: Castle et al., 2015c)
indicator saturation for discovering outliers and breaks, as well as selecting over other variables. Although
knowing the in-sample DGP need not suffice for successful forecasting after out-of-sample shifts like fi-
nancial crises and pandemics, ‘robust variants’ of selected models can then avoid systematic mis-forecasts
(see, for example, Martinez et al., 2019; Doornik et al., 2020a).
Saturation estimation has approximately K = k+2n candidates for k regressors, with 2K possible models,
requiring selection with more candidate variables, K, than observations, n. Selection criteria like AIC
(Akaike, 1973), BIC (Schwarz, 1978), and HQ (Hannan and Quinn, 1979) are insufficient in this setting.
For saturation estimation, we select at a tight significance level, α = 1/K, retaining subject-theory variables
and other regressors. When forecasting is the aim, analyses and simulations suggest loose significance
for then selecting other regressors, close to the 10% to 16% implied significance level of AIC, regardless
of location shifts at or near the forecast origin (Castle et al., 2018). At loose levels, Autometrics can find
multiple undominated terminal models across paths, and averaging over these, a univariate method and a
robust forecasting device can be beneficial, matching commonly found empirical outcomes. The approach
applies to systems, whence selection significance of both indicators and variables is judged at the system
level. Capturing in-sample location shifts remains essential (Doornik et al., 2020b).
There are costs and benefits of selection for forecasting. Selection at a loose significance level implies
excluding fewer relevant variables that contribute to forecast accuracy, but retaining more irrelevant vari-
ables that are adventitiously significant, although fewer than by simply averaging over all sub-models
(Hoeting et al., 1999). Retaining irrelevant variables that are subject to location shifts worsens forecast
performance, but their coefficient estimates are driven towards zero when updating estimates as the fore-
cast origin moves forward. Lacking omniscience about future values of regressors that then need to be
forecast, not knowing the DGP need not be costly relative to selecting from a general model that nests it

34 This subsection was written by David F. Hendry.

39
(Castle et al., 2018). Overall, the costs of model selection for forecasting are small compared to the more
fundamental benefit of finding location shifts that would otherwise induce systematic forecast failure.

2.4.5. Cross-validation for time-series data35

When building a predictive model, its purpose is usually not to predict well the already known sam-
ples, but to obtain a model that will generalise well to new, unseen data. To assess the out-of-sample
performance of a predictive model we use a test set that consists of data not used to estimate the model.
However, as we are now using only parts of the data for model building, and other parts of the data for
model evaluation, we are not making the best possible use of our data set, which is a problem especially if
the amount of data available is limited. Cross-validation (CV), first introduced by Stone (1974), is a widely
used standard technique to overcome this problem (Hastie et al., 2009) by using all available data points
for both model building and testing, therewith enabling a more precise estimation of the generalisation
error and allowing for better model selection. The main idea of (k-fold) cross-validation is to partition
the data set randomly into k subsets, and then use each of the k subsets to evaluate a model that has been
estimated on the remaining subsets. An excellent overview of different cross-validation techniques is given
by Arlot and Celisse (2010).
Despite its popularity in many fields, the use of CV in time series forecasting is not straightforward.
Time series are often non-stationary and have serial dependencies. Also, many forecasting techniques
iterate through the time series and therewith have difficulties dealing with missing values (withheld for
testing). Finally, using future values to predict data from the past is not in accordance with the normal
use case and therewith seems intuitively problematic. Thus, practitioners often resort to out-of-sample
evaluation, using a subset from the very end of a series exclusively for evaluation, and therewith falling
back to a situation where the data are not used optimally.
To overcome these problems, the so-called time-series cross-validation (Hyndman and Athanasopoulos,
2018) extends the out-of-sample approach from a fixed origin to a rolling origin evaluation (Tashman,
2000). Data is subsequently moved from the out-of-sample block from the end of the series to the training
set. Then, the model can be used (with or without parameter re-estimation) with the newly available
data. The model re-estimation can be done on sliding windows with fixed window sizes or on expanding
windows that always start at the beginning of the series (Bell and Smyl, 2018).
However, these approaches extending the out-of-sample procedure make again not optimal use of the
data and may not be applicable when only small amounts of data are available. Adapting the original CV
procedure, to overcome problems with serial correlations, blocked CV approaches have been proposed in
the literature, where the folds are chosen in blocks (Racine, 2000; Bergmeir and Benı́tez, 2012) and/or data
around the points used for testing are omitted (Racine, 2000; Burman et al., 1994). Finally, it has been
shown that with purely autoregressive models, CV can be used without modifications, i.e., with randomly
choosing the folds (Bergmeir et al., 2018). Here, CV estimates the generalisation error accurately, as long as
the model errors from the in-sample fit are uncorrelated. This especially holds when models are overfitting.
Underfitting can be easily detected by checking the residuals for serial correlation, e.g., with a Ljung-Box
test (Ljung and Box, 1978). This procedure is implemented in the forecast package (Hyndman et al., 2020)
in R (R Core Team, 2020), in the function CVar.

35 This subsection was written by Christoph Bergmeir.

40
2.5. Combining forecasts
2.5.1. Forecast combination: a brief review of statistical approaches36
Given N forecasts of the same event, forecast combination involves estimation of so called combination
weights assigned to each forecast, such that the accuracy of the combined forecast generally outperforms
the accuracy of the forecasts included. Early statistical approaches adopted a range of strategies to estimate
combination weights including (i) minimizing in-sample forecast error variance among forecast candidates
(Bates and Granger, 1969; Newbold and Granger, 1974; Min and Zellner, 1993), (ii) formulation and es-
timation via ordinary least squares regression (Granger and Ramanathan, 1984; MacDonald and Marsh,
1994), (iii) use of approaches based on Bayesian probability theory (Bunn, 1975; Bordley, 1982; Clemen
and Winkler, 1986; Diebold and Pauly, 1990; Raftery, 1993), (iv) and the use of regime switching and time
varying weights recognising that weights can change over time (Diebold and Pauly, 1987; Elliott et al.,
2005; Lütkepohl, 2011; Tian and Anderson, 2014). Barrow and Kourentzes (2016) contains a very good
documentation and empirical evaluation of a range of these early approaches, while De Menezes et al.
(2000) and Armstrong (2001b) contain guidelines on their use.
Recent statistical approaches use a variety of techniques to generate forecasts and/or derive weights.
Kolassa (2011) apply so called Akaike weights based on Akaike Information Criterion (Sakamoto et al.,
1986), while bootstrapping has been used to generate and combine forecast from exponential smoothing
(Cordeiro and Neves, 2009; Barrow et al., 2020; Bergmeir et al., 2016), artificial neural networks (Barrow
and Crone, 2016a,b), and other forecast methods (Athanasopoulos et al., 2018; Hillebrand and Medeiros,
2010; Inoue and Kilian, 2008). Barrow and Crone (2016b) developed cross-validation and aggregating
(Crogging) using cross-validation to generate and average multiple forecasts, while more recently, combi-
nations of forecasts generated from multiple temporal levels has become popular (Kourentzes et al., 2014;
Athanasopoulos et al., 2017; Kourentzes and Athanasopoulos, 2019). These newer approaches recognise
the importance of forecast generation in terms of uncertainty reduction (Petropoulos et al., 2018a), the
creation of diverse forecasts (Brown et al., 2005a; Lemke and Gabrys, 2010), and the pooling of forecasts
(Kourentzes et al., 2019; Lichtendahl Jr and Winkler, 2020).
Now nearly 50 years on from the seminal work of Bates and Granger (1969), the evidence that statisti-
cal combinations of forecasts improves forecasting accuracy is near unanimous, including evidence from
competitions (Makridakis et al., 1982; Makridakis and Hibon, 2000; Makridakis et al., 2020c), and empir-
ical studies (Elliott et al., 2005; Jose and Winkler, 2008; Andrawis et al., 2011; Kourentzes et al., 2019).
Still, researchers have tried to understand why and when combinations improve forecast accuracy (Palm
and Zellner, 1992; Petropoulos et al., 2018a; Timmermann, 2006), and the popularity of the simple aver-
age (Chan and Pauwels, 2018; Smith and Wallis, 2009a; Claeskens et al., 2016). Others have investigated
properties of the distribution of the forecast error beyond accuracy considering issues such as normality,
variance, and in out-of-sample performance of relevance to decision making (Makridakis and Winkler,
1989; Chan et al., 1999; Barrow and Kourentzes, 2016).
Looking forward, evidence suggests that the future lies in the combination of statistical and machine
learning generated forecasts (Makridakis et al., 2020c), and in the inclusion of human judgement (Gupta,
1994; Wang and Petropoulos, 2016; Petropoulos et al., 2018b). Additionally, there is need to investigate

36 This subsection was written by Devon K. Barrow.

41
such issues as decomposing combination accuracy gains, constructing prediction intervals (Koenker, 2005;
Grushka-Cockayne and Jose, 2020), and generating combined probability forecasts (Raftery et al., 1997;
Ranjan and Gneiting, 2010; Hall and Mitchell, 2007; Clements and Harvey, 2011). Finally, there is need
for the results of combined forecasts to be more interpretable and suitable for decision making (Bordignon
et al., 2013; Graefe et al., 2014; Barrow and Kourentzes, 2016; Todini, 2018).

2.5.2. Density forecast combinations37

Density forecasts provide an estimate of the future probability distribution of a random variable of in-
terest. Unlike point forecasts (and point forecasts supplemented by prediction intervals) density forecasts
provide a complete measure of forecast uncertainty. This is particularly important as it allows decision
makers to have full information about the risks of relying on the forecasts of a specific model. Policy mak-
ers like the Bank of England, the European Central Bank and Federal Reserve Banks in the US routinely
publish density forecasts of different macroeconomic variables such as inflation, unemployment rate and
GDP. In finance, density forecasts find application, in particular, in the areas of financial risk management
and forecasting of stock returns (see, for example, Tay and Wallis, 2000; Berkowitz, 2001; Guidolin and
Timmermann, 2006; Shackleton et al., 2010, inter alia).
Initial work on forecast combination focused on the combination of point forecasts. This literature
goes back to Bates and Granger (1969) and a summary of more recent work is provided in (Timmermann,
2006), while the ability of such methods to improve prediction accuracy is extensively documented (see,
for example, Stock and Watson, 2004; Hansen, 2008; Guidolin and Timmermann, 2009, inter alia). In
recent years attention has shifted towards evaluation, comparison and combination of density forecasts
with empirical applications that are mostly encountered in the areas of macroeconomics and finance. The
improved performance of combined density forecasts stems from the fact that pooling of forecasts allows
to mitigate potential misspecifications of the individual densities when the true population density is non-
normal. Combining normal densities yields a flexible mixture of normals density which can accommodate
heavier tails (and hence skewness and kurtosis), as well as approximate non-linear specifications (Hall and
Mitchell, 2007; Jore et al., 2010).
The predictive density combination schemes vary across studies and range from simple averaging of
individual density forecasts to complex approaches that allow for time-variation in the weights of predic-
tion models, also called experts (see Aastveit et al., 2019, for a comprehensive survey of density forecast
combination methods). A popular approach is to combine density forecasts using a convex combination of
experts’ predictions, so called ‘linear opinion pools’ (see, for example, Hall and Mitchell, 2007; Kascha and
Ravazzolo, 2010; Geweke and Amisano, 2011). In order to determine the optimal combination weights this
method relies on minimising Kullback-Leibler divergence of the true density from the combined density
forecast. These linear approaches have been extended by Gneiting and Ranjan (2013) to allow non-linear
transformations of the aggregation scheme and by Kapetanios et al. (2015), whose ‘generalised pools’ allow
the combination weights to depend on the (forecast) value of the variable of interest.
Billio et al. (2013) developed a combination scheme that allows the weights associated with each predic-
tive density to be time-varying, and propose a general state space representation of predictive densities and
combination schemes. The constraint that the combination weights must be non-negative and sum to unity

37 This subsection was written by Alisa Yusupova.

42
implies that linear and Gaussian state-space models cannot be used for inference and instead Sequential
Monte Carlo methods (particle filters) are required. More recently, McAlinn and West (2019) developed a
general Bayesian formulation for estimating time-varying linear (instead of convex) combinations of pre-
dictive densities. The method implemented in this work uses a pool of dynamic linear models to generate
individual predictive densities, and then combines these through another dynamic linear model.

2.5.3. Ensembles and predictive probability post processors38

Improved rational decisions are the final objective of modelling and forecasting. Relatively easy deci-
sions among a number of alternatives with predefined and known outcomes become hard when they are
conditioned by future unknown events. This is why one resorts to modelling and forecasting, but this is
insufficient. To be successful, one must account for the future conditioning event uncertainty to be in-
corporated into the decision-making process using appropriate Bayesian approaches, as described by the
decision theory literature (Berger, 1985; Bernardo, 1994; DeGroot, 2004). The reason is that taking a deci-
sion purely based on model forecasts is equivalent to assuming the future event (very unlikely, as we know)
to equal the forecasted value. Therefore, the estimation of the predictive probability density is the essen-
tial prerequisite to estimating the expected value of benefits (or of losses) to be compared and traded-off
in the decision-making process (Draper and Krnjajić, 2013). This highly increases the expected advantages
together with the likelihood of success and the robustness of the decision (Todini, 2017, 2018).
In the past, the assessment of the prediction uncertainty was limited to the evaluation of the confidence
limits meant to describe the quality of the forecast. This was done using continuous predictive densities, as
in the case of the linear regression, or more frequently in the form of predictive ensembles. These proba-
bilistic predictions, describing the uncertainty of the model forecasts given (knowing) the observations can
be used within the historical period to assess the quality of the models (Todini, 2016). When predicting
into the future, observations are no more available and what we look for (known as predictive probability)
is the probability of occurrence of the unknown “future observations” given (knowing) the model fore-
casts. This can be obtained via Bayesian inversion, which is the basis of several uncertainty post-processors
used in economy (Diebold et al., 1998), hydrology (Krzysztofowicz, 1999; Todini, 1999, 2008; Schwanen-
berg et al., 2015), meteorology (Granger and Pesaran, 2000; Katz and Lazo, 2011; Economou et al., 2016;
Reggiani and Boyko, 2019), etc. Accordingly, one can derive a predictive probability from a single model
forecast to estimate the expected value of a decision by integrating over the entire domain of existence all
the possible future outcomes and theirs effects, weighted with their appropriate probability of occurrence.
When several forecasts are available to a decision maker, the problem of deciding on which of them one
should rely upon becomes significant. It is generally hard to choose among several forecasts because one
model could be best under certain circumstances but rather poor under others. Accordingly, to improve
the available knowledge on a future unknown event, predictive densities are extended to multiple forecasts
to provide decision makers with a single predictive probability, conditional upon several model’s forecasts
(Raftery et al., 1997; Coccia and Todini, 2011).
A number of available uncertainty post processors can cope with multi-model approaches, such as
Bayesian model averaging (Raftery, 1993; Raftery et al., 1997), model output statistics (Glahn and Lowry,

38 This subsection was written by Ezio Todini.

43
1972; Wilkd, 2005), ensemble model output statistics (Gneiting et al., 2005), and model conditional pro-
cessor (Todini, 2008; Coccia and Todini, 2011).
Finally, important questions such as: (i) “what is the probability that an event will happen within the
next x hours?” and (ii) “at which time interval it will most likely occur?” can be answered using the multi-
temporal approach (Krzysztofowicz, 2014; Coccia, 2011) and results of its applications were presented in
Coccia (2011), Todini (2017), and Barbetta et al. (2017).

2.5.4. The wisdom of crowds39

Multiple experts’ forecasts are collected in a wide variety of situations: medical diagnostics, weather
prediction, forecasting the path of a hurricane, predicting the outcome of an election, macroeconomic fore-
casting, and more. One of the central findings from the forecasting literature is that there is tremendous
power in combining such experts’ forecasts into a single forecast. The simple average, or what Surowiecki
refers to as ‘the wisdom of crowds’ (Surowiecki, 2005), has been shown to be a surprisingly robust com-
bined forecast in the case of point forecasting (Clemen and Winkler, 1986; Clemen, 1989; Armstrong,
2001a). The average forecast is more accurate than choosing a forecast from the crowd at random and is
sometimes even more accurate than nearly all individuals (Mannes et al., 2012). The average point fore-
cast also often outperforms more complicated point aggregation schemes, such as weighted combinations
(Smith and Wallis, 2009b; Soule et al., 2020).
Mannes et al. (2012) highlight two crucial factors that influence the quality of the average point forecast:
individual expertise and the crowd’s diversity. Of the two: “The benefits of diversity are so strong that one
can combine the judgments from individuals who differ a great deal in their individual accuracy and still
gain from averaging” (Mannes et al., 2012, page 234).
Larrick and Soll (2006) define the idea of ‘bracketing’: In the case of averaging, two experts can either
bracket the realisation (the truth) or not. When their estimates bracket, the forecast generated by taking
their average performs better than choosing one of the two experts at random; when the estimates do not
bracket, averaging performs equally as well as the average expert. Thus, averaging can do no worse than
the average expert, and with some bracketing, it can do much better. Modern machine learning algorithms
such as the random forest exploit this property by averaging forecasts from hundreds of diverse experts
(here, each “expert” is a regression tree; Grushka-Cockayne et al., 2017a).
Only when the crowd of forecasts being combined has a high degree of dispersion in expertise, some
individuals in the crowd might stand out, and in such cases, there could be some benefits to chasing a
single expert forecaster instead of relying on the entire crowd. Mannes et al. (2014) suggest that combining
a small crowd can be especially powerful in practice, offering some diversity among a subset of forecasters
with an minimum level of expertise.
When working with probabilistic forecasting, averaging probabilities is the most widely used probabil-
ity combination method Cooke (1991); Hora (2004); Clemen (2008). Stone (1961) labelled such an average
the linear opinion pool. O’Hagan et al. (2006) claimed that the linear opinion pool is: “hard to beat in
practice”.
Although diversity benefits the average point forecast, it can negatively impact the average probability
forecast. As the crowd’s diversity increases, the average probability forecast becomes more spread out,

39 This subsection was written by Yael Grushka-Cockayne.

44
or more underconfident (Dawid et al., 1995; Hora, 2004; Ranjan and Gneiting, 2010). Averaging quan-
tiles, instead of probabilities, can offer sharper and better calibrated forecasts (Lichtendahl et al., 2013).
Trimmed opinion pools can be applied to probability forecasts, also resulting in better calibrated forecasts
(Jose et al., 2014).
The ubiquity of data and the increased sophistication of forecasting methods results in more use of
probabilistic forecasts. While probabilities are more complex to elicit, evaluate, and aggregate compared
to point estimates, they do contain richer information about the uncertainty of interest. The wisdom of
combining probabilities, however, utilises diversity and expertise differently than combining point fore-
casts. When relying on a crowd, eliciting point forecasts versus eliciting probabilities can significantly
influence the type of aggregation one might choose to use.

2.6. Data-driven methods

2.6.1. Forecasting with big data40
Are advances in computing and data availability going to pave the way for a revolution in forecasting?
Is big data going to significantly increase the forecast accuracy of macroeconomic forecasts? The last two
decades have seen a proliferation of literature on forecasting using big data (Varian, 2014; Swanson and
Xiong, 2018; Hassani and Silva, 2015). But the jury is still out as to whether the promised improvements
in forecast accuracy can be realised for macroeconomic phenomena.
The tools used to analyse big data focus on regularisation techniques to achieve dimension reduction,
see Kim and Swanson (2014, 2018) for a summary of the literature. This can be achieved through selection
(such as Autometrics, Doornik, 2018), shrinkage (including ridge regression: Hoerl and Kennard, 1970;
Lasso: Tibshirani, 1996; Zou, 2006; and Elastic Nets: Zou and Hastie, 2005), variable combination (such
as Principal Components Analysis and Partial Least Squares: Bai and Ng, 2008; Bai and Ng, 2009; Stock
and Watson, 2012; Ding and Hwang, 1999; Forni et al., 2000; Zou et al., 2006), and machine learning
methods (including Artificial Neural Networks: Zhang et al., 1998; Medeiros et al., 2006). Many of these
methods are ‘black boxes’ where the algorithms are not easily interpretable, and so they are mostly used
for forecasting rather than for policy analysis.
Machine learning methods are seen as an efficient approach to dealing with big data sets (Athey, 2018).
Big data has been effectively used to estimate the forecast origin which, if improved, would lead to better
forecasts absent any later shifts. Nowcasting can benefit from large data sets as the events have happened
and the information is available; see Castle et al. (2017) for a nowcasting application. However, the benefits
of big data are not as evident in a forecasting context where the future values of all added variables also
need to be forecast and are as uncertain as the variable of interest.
Macroeconomic time series data are highly non-stationary, with stochastic trends and structural breaks.
Regularisation methods (such as cross-validation and hold-back) often assume that the data generating
process does not change over time. Forecasting models that assume the data are drawn from a stationary
distribution (even after differencing) do not forecast well ex ante. So while there seems to be lots of mileage
in improving forecasts using big data, as they allow for more flexible models that nest wider information
sets, more general dynamics and many forms of non-linearity, the statistical problems facing ‘small’ data

40 This subsection was written by Jennifer L. Castle.

45
forecasting models do not disappear (Harford, 2014; Doornik and Hendry, 2015). Castle et al. (2020a) do
not find improvements in forecasting from big data sets over small models. It is essential to keep in mind
the classical statistical problems of mistaking correlation for causation, ignoring sampling biases, finding
excess numbers of false positives and not handling structural breaks and non-constancies both in- and
out-of-sample, in order to guard against these issues in a data abundant environment.

2.6.2. Forecasting on distributed systems41

Big data is normally accompanied by the nature that observations are indexed by timestamps, giving
rise to big data time series characterised by high frequency and long-time span. Processing big data time
series is obstructed by a wide variety of complications, such as significant storage requirements, algo-
rithms’ complexity and high computational cost (L’heureux et al., 2017; Wang et al., 2018a; Galicia et al.,
2018; Wang et al., 2020b). These limitations accelerate the great demand for scalable algorithms. Nowa-
days, increasing attention has been paid to developing data mining techniques on distributed systems for
handling big data time series, including but not limited to processing (Mirko and Kantelhardt, 2013), de-
composition (Bendre and Manthalkar, 2019), clustering (Ding et al., 2015), classification (Triguero et al.,
2015), and forecasting (Galicia et al., 2018).
Distributed systems, initially designed for independent jobs, do not support to deal with dependen-
cies among observations, which is a critical obstacle in time series processing (Li et al., 2014; Wang et al.,
2020b). Various databases (e.g., InfluxDB42 , OpenTSDB43 , RRDtool44 , and Timely45 ) can function as stor-
age platforms for time series data. However, none of these databases supports advanced analysis, such as
modelling, machine learning algorithms and forecasting. Additional considerations are therefore required
in further processing time series on distributed systems. Mirko and Kantelhardt (2013) developed the
Hadoop.TS library for processing large-scale time series by creating a time series bucket. Li et al. (2014)
designed an index pool serving as a data structure for assigning index keys to time series entries, allowing
time series data to be sequentially stored on HDFS (Hadoop Distributed File System: Shvachko et al., 2010)
for MapReduce (Dean and Ghemawat, 2008) jobs. Chen et al. (2019a) proposed a data compression and
abstraction method for large-scale time series to facilitate the periodicity-based time series prediction in a
parallel manner.
The evolution of the algorithms for efficiently forecasting big data time series on distributed systems
is largely motivated by a wide range of applications including meteorology, energy, finance, transporta-
tion and farming (Galicia et al., 2018; Chen et al., 2019a; Hong and Pinson, 2019; Sommer et al., 2020).
Researchers have made several attempts to make machine learning techniques available for big data time
series forecasting on distributed systems (Li et al., 2014; Talavera-Llames et al., 2016; Galicia et al., 2019;
Xu et al., 2020). Talavera-Llames et al. (2016) presented a nearest neighbours-based algorithm imple-
mented for Apache Spark (Zaharia et al., 2016) and achieved satisfactory forecasting performance. Galicia
et al. (2018) proposed a scalable methodology which enables Spark’s MLlib (Meng et al., 2016) library to

41 This subsection was written by Xiaoqian Wang.

42 Available at https://www.influxdata.com/time-series-database/
43 Available at http://opentsdb.net/
44 Available at https://oss.oetiker.ch/rrdtool/
45 Available at https://code.nsa.gov/timely/

46
conduct multi-step forecasting by splitting the multi-step forecasting problem into h sub-problems (h is
the forecast horizon).
Another strand of the literature on forecasting big data time series is to improve time-consuming es-
timation methods using a MapReduce framework. Sheng et al. (2013) learned the parameters of echo
state networks for time series forecasting by designing a parallelised extended Kalman filter involving
two MapReduce procedures. Recently, Sommer et al. (2020) accurately estimated coefficients of a high-
dimensional ARX model by designing two online distributed learning algorithms. Wang et al. (2020b)
resolved challenges associated with forecasting ultra-long time series from a new perspective that global
estimators are approximated by combining the local estimators obtained from subseries by minimising a
global loss function. Besides, inspired by the no-free-lunch theorem (Wolpert and Macready, 1997), model
selection and model combination are involved in finalisation of algorithms for forecasting on distributed
systems (e.g., Li et al., 2014; Galicia et al., 2019; Bendre and Manthalkar, 2019; Xu et al., 2020).

2.6.3. Agent-based models46

Time series forecasting involves use of historical data to predict values for a specific period time in fu-
ture. This approach assumes that recent and historical patterns in the data will continue in the future. This
assumption is overly ingenuous. However, this is not reliable in some situations. For example, (i) forecast-
ing COVID-19 cases where, due to interventions and control measures taken by the governments and due
to the change in personal behaviour, the disease transmission pattern changes rapidly, and (ii) forecasting
sales of a new product: external factors such as advertisement, promotions, social learning, and imitation
of other individuals change the system behaviour.
In such circumstances to make reliable forecasts it is important to take into account all information that
might influence the variable that is being forecast. This information includes a variety of environmental-
level and individual-level factors. An agent-based modelling is a powerful tool to explore such complex
systems. Agent-based modelling approach is useful when, (ii) data availability is limited, (ii) uncertainty of
various interventions in place and a rapidly changing social environment, and (iii) limited understanding
of the dynamics of the variable of interest.
Agent-based modelling disaggregates systems into individual level and explores the aggregate impact
of individual behavioural changes on the system as a whole. In other words, the key feature of agent-
based modelling is the bottom-up approach to understand how a system’s complexity arises, starting with
individual level. As opposed to this, the conventional time series forecasting approaches are considered
top-down approaches.
Agent-based models have two main components: (i) Agents, and (ii) Rules, sometimes referred as pro-
cedures and interactions. Agents are individuals with autonomous behaviour. Agents are heterogeneous.
Each agent individually assesses on the basis of a set of rules. An agent-based modelling approach simu-
lates how heterogeneous agents interact and behave to assess the role of different activities on the target
variable.
A general framework for Agent-based modelling involves three main stages (See Figure 3): (i) setup en-
vironments and agents, (ii) agent-based modelling, and (iii) calibration and validation. The first two steps

46 This subsection was written by Thiyanga S. Talagala.

47
are self-explanatory. The final step involves calibration of the model with empirical data and then eval-
uates whether the agent-based model mirrors the real-world system/target. The validation step involves
testing the significance of the difference between agent-based model results and real data collected about
the target. One of the main challenges in designing an agent-based model is finding a balance between
model simplicity and model realism. The KISS principle (keep it simple, stupid), introduced by Axelrod
(1997) is often cited as an effective strategy in agent-based modelling. A high level of expertise in the area
of the subject is necessary when developing an agent-based model.

Target
(Calibration and Validation)

Agent-based Modelling
Bottom-up approach

Setup agents and Environment

environment
Interaction

Agent’s Agent’s Agent’s

behaviour behaviour …. behaviour

Agent 1 Agent 2 …. Agent n

Figure 3: Framework for Agent-based modelling.

Despite these limitations and challenges, agent-based modelling has been used extensively to model
infectious disease transmission and forecasting (Tracy et al., 2018; Venkatramanan et al., 2018). Agent-
based modelling approaches have been widely used in early phases of the COVID-19 outbreak, to assess
the impact of different interventions on disease spread and forecasts (Wallentin et al., 2020). In a review
paper, Weron (2014) states some applications of agent-based models for electricity demand forecasting.
Xiao and Han (2016) use agent-based models to forecast new product diffusion. Furthermore, thinking the
other way around, Hassan et al. (2013) explain how forecasting principles can be applied in agent-based
modelling.

2.6.4. Feature-based time series forecasting47

A time series feature can be any statistical representation of time series characteristics. A vast majority
of time series mining tasks are based on similarity quantification using their feature representations, in-
cluding but not limited to time series clustering (Wang et al., 2006; Kang et al., 2014, 2015), classification
(Nanopoulos et al., 2001; Fulcher et al., 2013), anomaly detection (Kang, 2012; Talagala et al., 2020a), and
forecasting (Kang et al., 2017; Montero-Manso et al., 2020). The choice of features depends on the nature of

47 This subsection was written by Yanfei Kang.

48
the data and the application context. The state-of-the-art time series feature representation methods quan-
tify a wide range of time series characteristics, including simple summary statistics, stationarity (Montero-
Manso et al., 2020; Wang et al., 2020c), model fits (Fulcher and Jones, 2014; Christ et al., 2018), time series
imaging (Li et al., 2020f), and others. In the forecasting community, two lines of forecasting approaches
have been developed using time series features, namely feature-based model selection and feature-based
model combination. The motivation behind them is no single model always performs the best for all time
series. Instead of choosing one model for all the data, features can be used to obtain the most appropriate
model or the optimal combination of candidate models, per series.
As early as in 1972, Reid (1972) argues that time series characteristics provide valuable information in
forecast model selection, which is further echoed by Makridakis and Hibon (1979). One way to forecast
an extensive collection of time series is to select the most appropriate method per series according to its
features. Pioneer studies focus on rule-based methods (for example, Arinze, 1994; Wang et al., 2009b)
to recommend the “best” forecasting model per series based on its features. Another line of approaches
apply regression to study how useful features are in predicting which forecasting method performs best
(for example, Meade, 2000; Petropoulos et al., 2014). With the advancement of machine learning, more
recent literature uses “meta-learning” to describe the process of automatically acquiring knowledge for
forecast model selection. The first such study is by Prudêncio and Ludermir (2004), who apply decision
trees for forecast model selection. Lemke and Gabrys (2010) compare different meta-learning approaches
to investigate which model works best in which situation. Kang et al. (2017) propose using feature spaces to
visualise the strengths and weaknesses of different forecasting methods. Other algorithms such as neural
networks and random forecasts are also applied to forecast model selection (Kück et al., 2016; Talagala
et al., 2018).
One of the pioneering studies in feature-based forecast combination is the rule-based approach by
Collopy and Armstrong (1992), who develop 99 rules for forecast combination based on 18 features. Re-
cently, Kang et al. (2020) use 26 features to predict the performances of nine forecasting methods with
nonlinear regression models, and obtain the combination weights as a tailored softmax function of the
predicted forecasting errors. The feature-based forecast model averaging (FFORMA) framework proposed
by Montero-Manso et al. (2020) employ 42 features to estimate the optimal combination weights of nine
forecasting methods based on extreme gradient boosting (XGBoost, Chen and Guestrin, 2016). Li et al.
(2020f) first transform time series into images, and use features extracted from images to estimate the opti-
mal combination weights. For feature-based interval forecasting, Wang et al. (2020c) investigate how time
series features affect the relative performances of prediction intervals from different methods, and propose
a general feature-based interval forecasting framework to provide reliable forecasts and their uncertainty
estimation.

2.6.5. Forecasting with bootstrap48

Bootstrap methodology has been widely applied in many areas of research, including time series analy-
sis. The bootstrap procedure (Efron, 1979) is a very popular methodology for independent data because of
its simplicity and nice properties. It is a computer-intensive method that presents solutions in situations
where the traditional methods fail or are very difficult to apply. However, Efron’s bootstrap (iid bootstrap)

48 This subsection was written by Clara Cordeiro.

49
has revealed itself inefficient in the context of dependent data, such as in the case of time series, where the
dependence structure arrangement has to be kept during the resampling scheme.
Most of the resampling for dependent data consider segments of the data to define blocks, such that
the dependence structure within each block can be kept. Different versions of blocking differ in the way as
blocks are constructed: the nonoverlapping block bootstrap (Carlstein, 1990), the moving block bootstrap
(Künsch, 1989), the circular block bootstrap (Politis and Romano, 1992), and the stationary block bootstrap
(Politis and Romano, 1994). But, if the time series process is driven from iid innovations, another way of
resampling can be used.
The iid Bootstrap can then be easily extended to a dependent setup. That was the spirit of sieve boot-
strap proposed by Bühlmann (1997). This method is based on the idea of fitting parametric models first
and resampling from the residuals. Such models include, for example, the linear regression (Freedman,
1981) and autoregressive time series (Efron and Tibshirani, 1986). This approach is different from the pre-
vious bootstrap methods for dependent data; the sample bootstrap is (conditionally) stationary and does
not present a structure of dependence. Another different feature is that the sieve bootstrap sample is not a
subsample from the original data, as in the previous methods. Observe that even if the sieve bootstrap is
based on a parametric model, it is nonparametric in its spirit. The AR model here is just used to filter the
residuals series.
A few years ago, the sieve bootstrap was used for estimating forecast intervals (Zagdański, 2001; Andrés
et al., 2002). Motivated by these works, Cordeiro and Neves (2006, 2009, 2010) developed a procedure
to estimate point forecasts. The idea of these authors was to fit a model to the time series, extract the
residuals and then apply the sieve bootstrap to the residuals. Further developments of this procedure
include the estimation of forecast intervals (Cordeiro and Neves, 2014) and also the detection, estimation
and imputation of missing data (Cordeiro and Neves, 2013). In a recent work (Bergmeir et al., 2016)
a similar approach was also consider, the residuals were extracted and resampled using moving block
bootstrap.
Bickel and Freedman (1981) and later in Angus (1992) showed that in extreme value theory, the boot-
strap version for the maximum (or minimum) does not converge to the extremal limit laws. Zelterman
(1993) pointed out that “to resample the data for approximating the distribution of the k largest obser-
vations would not work because the ‘pseudo-samples’ would never have values greater than Xn:n ”49 . A
method considering to resample a smaller size than the original sample was proposed in Hall (1990). Re-
cently, Neves and Cordeiro (2020) used this idea and developed a preliminary work in modelling and
forecasting extremes in time series.

2.6.6. Bagging for time series forecasting50

The term bagging was proposed by Breiman (1996) to describe the generation of several versions of
a predictor, via Bootstrap procedures, with a final aggregation stage. Thus, “bootstrap aggregating” was
established as bagging. The main idea is to improve predictors’ accuracy once the data sets, draw randomly
with replacement, will approximating the original distribution. The author argues that bagging works for
unstable procedures, but it was not tested for time series. Years after, Kilian and Inoue (2004) suggested

49 max(X , · · · , X ).
1 n
50 This subsection was written by Fernando Luiz Cyrino Oliveira.

50
the first attempts for temporal dependent data. For data-driven methods, to forecasting and simulation
time series and deal with predictors ensembles, bagging has shown as a powerful tool.
A general framework for ensemble forecasting methods involves four main stages: (i) data treatment,
(ii) resampling, (iii) forecasting, and (iv) aggregation. However, for time series, bootstrap should be done
carefully, as the serial dependence and non-stationarity must be considered.
This led Bergmeir et al. (2016) to propose a bagging version for exponential smoothing, the Bagged
ETS. As pre-treatment, after a Box-Cox transformation, the series is decomposed into trend, seasonal, and
remainder components via STL decomposition (Cleveland et al., 1990). The resampling stage uses moving
block bootstrap (MBB: Lahiri and Lahiri, 2003), applied to the remainder. There are several discussions in
the literature about this procedure, mainly regarding the size of the blocks. MBB resampling the collection
of overlapping (consecutive) blocks of observations. The idea is to keep the structure still present in the
remainder. The forecasts are obtained via ETS methods and, for the final aggregation, the authors adopted
the median. Their method is evaluated on the M3 data set and outperformed the original benchmarks. The
work of Bergmeir et al. (2016) inspired many others: Dantas et al. (2017) applied the idea for air transport
demand data and de Oliveira and Cyrino Oliveira (2018) for energy consumption, proposing the so-called
remainder sieve bootstrap (RSB).
Dantas and Cyrino Oliveira (2018) proposed an extension to the Bagged ETS where bagging and expo-
nential smoothing are combined with clustering methods. The approach aims to consider and reduce the
covariance effects among the ensemble time series, creating clusters of similar forecasts – since it could
impact the variance of the group. A variety of forecasts are selected from each cluster, producing groups
with reduced variance.
In light of the aforementioned, there are several possibilities for each stage of the mentioned framework.
In this context, to investigate the reasons why bagging works well for time series forecasting, Petropoulos
et al. (2018a) explored three sources of uncertainty: model form, data, and parameter. While arguably
bagging can handle all of them, Petropoulos et al. (2018a) showed that simply tackling model uncertainty
is enough for achieving a superior performance, leading to the proposal of a Bootstrap Model Combination
(BMC) approach, where different model forms are identified in the ensemble and fitted to the original data.
Finally, Meira et al. (2020) proposed “treating and pruning” strategies to improve the performance of
prediction intervals for both model selection and forecast combinations. Testing over a large set of real
time series from the M forecasting competitions, their results highlighted the importance of analysing the
prediction intervals of the ensemble series before the final aggregation.

2.6.7. Neural networks51

Neural Networks (NNs) or Artificial Neural Networks (ANNs) are mathematical formulations inspired
by the work and functioning of biological neurons. They are characterized by their ability to model non-
stationary, nonlinear and high complex datasets. This property along with the increased computational
power have put NNs in the frontline of research in most fields of science (De Gooijer and Hyndman, 2006;
Zhang et al., 1998).
A typical NN topology is consisted by three types of layers (input, hidden and output) and each layer is
consisted by nodes. The first layer in every NN, is the input layer and the number of its nodes corresponds

51 This subsection was written by Georgios Sermpinis.

51
to the number of explanatory variables (inputs). The last layer is the output layer and the number of nodes
corresponds to the number of response variables (forecasts). Between the input and the output layer, there
is one or more hidden layers where the nodes define the amount of complexity the model is capable of
fitting. Most NN topologies in the input and the first hidden layer contain an extra node, called the bias
node. The bias node has a fixed value of one and serves a function similar to the intercept in traditional
regression models. Each node in one layer has connections (weights) with all or a subset (for example, for
the convolutional neural network topology) of the nodes of the next layer.
NNs process the information as follows: the input nodes contain the explanatory variables. These vari-
ables are weighted by the connections between the input and the first hidden nodes, and the information
reaches to the hidden nodes as a weighted sum of the inputs. In the hidden nodes, there is usually a non-
linear function (such as the sigmoid or the RelU) which transform the information received. This process
is repeated until the information reaches the output layer as forecasts. NNs are trained by adjusting the
weights that connect the nodes in a way that the network maps the input value of the training data to the
corresponding output value. This mapping is based on a loss function, the choice of which depends on
the nature of the forecasting problem. The most common NN procedure, is the back-propagation of errors
(Shapiro, 2000).
The simpler and most common NN topology, is the Multilayer Forward Perceptron (MLP). In MLP,
the hidden nodes contain the sigmoid function and the information moves in forward direction (from the
inputs to the output nodes). An another NN topology where the information moves also only in a forward
direction is the Radial Basis Function NN (RBF). Now the hidden neurons compute the Euclidean distance
of the test case from the neuron’s centre point and then applies the Gaussian function to this distance using
the spread values. Recurrent Neural Networks (RNNs) are NN topologies that allow previous outputs to be
used as inputs while having hidden states. The information moves both forwards and backwards. RNNs
have short-term memory and inputs are taken potentially from all previous values. MLPs, RBFs and RNNs
are universal function approximators (Hornik, 1991; Schäfer and Zimmermann, 2006; Park and Sandberg,
1991). However, the amount of NN complexity in terms of hidden layers and nodes to reach this property,
might make the NN topology computationally unfeasible to train.

2.6.8. Deep probabilistic forecasting models52

Neural networks can be equipped to provide not only a single-valued forecast, but rather the entire
range of values possible in a number of ways. We will discuss three selected approaches in the following,
but remark that this is a subjective selection and is by far not comprehensive.53

1. Analogously to linear regression and Generalised Linear Models, obtaining probabilistic forecasts
can be achieved by the neural network outputting not the forecasted value itself but rather parame-
ters of probability distribution or density (Bishop, 2006). In forecasting, a prominent example is the
DeepAR model (Salinas et al., 2019b), which uses a recurrent neural network architecture and as-
sumes the probability distribution to be from a standard probability density function (e.g., negative

52 This subsection was written by Tim Januschowski.

53 For example, contemporary topics in machine learning such as generative adversarial networks can be naturally lifted to fore-
casting and similarly, more traditional probabilistic machine learning approaches such as Gaussian Processes (Maddix et al., 2018).
We ignore the important area of Bayesian deep learning (see Wang and Yeung, 2016, for a survey) entirely here for lack of space.

52
 binomial or Student’s t). Variations are possible, with either non-standard output distributions in
forecasting such as the multinomial distribution (Rabanser et al., 2020) or via representing the prob-
ability density as cumulative distribution function (Salinas et al., 2019a) or the quantile function
(Gasthaus et al., 2019).
2. An alternative approach is to apply concepts for quantile regression (Koenker, 2005) to neural net-
works, e.g., by making the neural network produce values for selected quantiles directly (Wen et al.,
2017).
3. It is possible to combine neural networks with existing probabilistic models. For example, neural
networks can parametrise state space models (Durbin and Koopman, 2012) as an example for another
class of approaches (Rangapuram et al., 2018), dynamic factor models (Geweke, 1977) with neural
networks (Wang et al., 2019b) or deep temporal point processes (Turkmen et al., 2019).

The appeals of using neural networks for point forecasts carry over to probabilistic forecasts, so we will
only comment on the elegance of modern neural network programming frameworks. To the forecasting
model builder, the availability of auto gradient computation, the integration with highly-tuned optimisa-
tion algorithms and scalability considerations built into the frameworks, means that the time from model
idea to experimental evaluation has never been shorter. In the examples above, we brushed over the need
to have loss functions with which we estimate the parameters of the neural networks. Standard negative
log-likelihood based approaches are easily expressible in code as are approaches based on non-standard
losses such as the continuous ranked probability score (Gneiting et al., 2007). With open source proliferat-
ing in the deep learning community, most of the above examples for obtaining probabilistic forecasts can
readily be test-driven (see, for example, Alexandrov et al., 2019).
For the future, we see a number of open challenges. Most of the approaches mentioned above are
univariate, in the following sense. If we are interested in forecasting values for all time series in a panel, we
may be interested in modelling the relationship among these time series. The aforementioned approaches
mostly assume independence of the time series. In recent years, a number of multivariate probabilistic
forecasting models have been proposed (Salinas et al., 2019a; Rangapuram et al., 2020), but much work
remains to obtain a better understanding. Another counter-intuitive challenge for neural networks is to
scale them down. Neural networks are highly parametrised, so in order to estimate parameters correctly,
panels with lots of time series are needed. However, a large part of the forecasting problem landscape
(Januschowski and Kolassa, 2019) consists of forecasting problems with only a few time series. Obtaining
good uncertainty estimates with neural networks in these settings is an open problem.

2.6.9. Machine learning54

Categorising forecasting methods into statistical and machine learning (ML) is not trivial, as various
criteria can be considered for performing this task (Januschowski et al., 2020). Nevertheless, more often
than not, forecasting methods are categorised as ML when then they do not prescribe the data-generating
process, e.g., through a set of equations, thus allowing for data relationships to be automatically learned
(Barker, 2020). In this respect, methods that build on unstructured, non-linear regression algorithms,
such as Neural Networks (NN), Decision Trees, Support Vector Machines, and Gaussian Processes, are
considered as ML (Makridakis et al., 2018).

54 This subsection was written by Evangelos Spiliotis.

53
 Since ML methods are data-driven, they are more generic and easier to be adapted to forecast series
of different characteristics (Spiliotis et al., 2020b). However, ML methods also display some limitations.
First, in order for ML methods to take full advantage of their capacity, sufficient data are required. Thus,
when series are relatively short and display complex patterns, such as seasonality and trend, ML meth-
ods are expected to provide sub-optimal forecasts if the data are not properly pre-processed (Zhang et al.,
1998; Makridakis et al., 2018). On the other hand, when dealing with long, high-frequency series, typi-
cally found in energy (Chae et al., 2016), stock market (Moghaddam et al., 2016), and demand (Carmo and
Rodrigues, 2004) related applications, ML methods can be applied with success. Second, computational in-
tensity may become relevant (Makridakis et al., 2020c), especially when forecasting numerous series at the
weekly and daily frequency (Seaman, 2018) or long-term accuracy improvements over traditional methods
are insignificant (Nikolopoulos and Petropoulos, 2018). Third, given that the effective implementation of
ML methods strongly depends on optimally determining the values of several hyper-parameters, related
both with the forecasting method itself and the training process, considerable complexity is introduced,
significant resources are required to set up the methods, and high experience and a strong background in
other fields than forecasting, such as programming and optimisation, are needed.
In order to deal with these limitations, ML methods can be applied in a cross-learning (CL) fashion
instead of a series-by-series one, i.e., allow the methods to learn from multiple series how to accurately
forecast the individual ones (Makridakis et al., 2020c). The key principle behind CL is that, although series
may differ, common patterns may occur among them, especially when data are structured in a hierarchical
way and additional information, such as categorical attributes and exogenous/explanatory variables, is
provided as input (Fry and Brundage, 2020). The CL approach has several advantages. First, computational
time can be significantly reduced as a single model can be used to forecast multiple series simultaneously.
Second, methods trained in a particular dataset can be effectively used to provide forecasts for series of
different datasets that display similar characteristics (transfer-learning), thus allowing the development
of generalised forecasting methods (Oreshkin et al., 2020a). Third, data limitations are mitigated and
valuable information can be exploited at global level, thus allowing for patterns shared among the series,
such as seasonal cycles (Dekker et al., 2004) and special events (Huber and Stuckenschmidt, 2020), to be
effectively captured.
Based on the above, CL is currently considered the most effective way of applying ML for batch time
series forecasting. Some state-of-the-art implementations of CL include long short-term memory NNs
(Smyl, 2020), deep NNs based on backward and forward residual links (Oreshkin et al., 2020b), feature-
based XGBoost (Montero-Manso et al., 2020), and gradient boosted decision trees (Bojer and Meldgaard,
2020).

2.6.10. Machine learning with (very) noisy data55

With the advent of big data, machine learning now plays a leading role in forecasting.56 There are two
primary reasons for this. First, conventional ordinary least squares (OLS) estimation is highly suscepti-
ble to overfitting in the presence of a large number of regressors (or features). OLS maximises the fit of

55 This subsection was written by David E. Rapach.

56 We do not make a sharp distinction between statistical learning and machine learning. For brevity, we use the latter throughout
this subsection.

54
the model over the estimation (or training) sample, which can lead to poor out-of-sample performance; in
essence, OLS over-responds to noise in the data, and the problem becomes magnified as the number of fea-
tures grows. A class of machine-learning techniques, which includes the popular least absolute shrinkage
and selection operator (LASSO, Tibshirani, 1996) and elastic net (ENet, Zou and Hastie, 2005), employs
penalised regression to improve out-of-sample performance with large numbers of features. The LASSO and
ENet guard against overfitting by shrinking the parameter estimates toward zero.
Very noisy data – data with a very low signal-to-noise ratio – exacerbate the overfitting problem. In
such an environment, it is vital to induce adequate shrinkage to guard against overfitting and more reliably
uncover the predictive signal amidst all the noise. For LASSO and ENet estimation, a promising strategy
is to employ a stringent information criterion, such as the Bayesian information criterion (BIC, Schwarz,
1978), to select (or tune) the regularisation parameter governing the degree of shrinkage (often denoted by
λ). Wang et al. (2009a) and Fan and Tang (2013) modify the BIC penalty to account for a diverging number
of features, while Hui et al. (2015) refine the BIC penalty to include the value of λ. These BIC variants
induce a greater degree of shrinkage by strengthening the BIC’s penalty term, making them useful for
implementing the LASSO and ENet in noisy data environments; see Filippou et al. (2020) for a recent
empirical application.
A second reason for the popularity of machine learning in the era of big data is the existence of powerful
tools for accommodating complex predictive relationships. In many contexts, a linear specification appears
overly restrictive, as it may neglect important nonlinearities in the data that can potentially be exploited to
improve forecasting performance. Neural networks (NNs) are perhaps the most popular machine-learning
device for modelling nonlinear predictive relationships with a large number of features. Under a reason-
able set of assumptions, a sufficiently complex NN can approximate any smooth function (for example,
Cybenko, 1989; Funahashi, 1989; Hornik et al., 1989; Barron, 1994).
By design, NNs are extremely flexible, and this flexibility means that a large number of parameters (or
weights) need to be estimated, which again raises concerns about overfitting, especially with very noisy
data. The weights of a NN are typically estimated via a stochastic gradient descent (SGD) algorithm, such
as Adam (Kingma and Ba, 2015). The SGD algorithm itself has some regularising properties, which can be
strengthened by adjusting the algorithm’s hyperparameters. We can further guard against overfitting by
imposing a dropout rate (Hinton et al., 2012; Srivastava et al., 2014) for the weights.
Perhaps the quintessential example of a noisy data environment is forecasting asset returns, especially
at short horizons (e.g., monthly). Because many asset markets are reasonably efficient, most of the fluc-
tuations in returns are inherently unpredictable – they reflect the arrival of new information, which, by
definition, is unpredictable. This does not mean that we should not bother trying to forecast returns, as
even a seemingly small degree of return predictability can be economically significant (for example, Camp-
bell and Thompson, 2008). Instead, it means that we need to be particularly mindful of overfitting when
forecasting returns in the era of big data.

2.6.11. Clustering-based forecasting57

The robustness of the forecasting process depends mainly on the characteristics of the target variable.
In cases of high nonlinear and volatile time series, a forecasting model may not be able to fully capture

57 This subsection was written by Ioannis Panapakidis.

55
and simulate the special characteristics, a fact that may lead to poor forecasting accuracy (Pradeepku-
mar and Ravi, 2017). Contemporary research has proposed some approaches to increase the forecasting
performance (Sardinha-Lourenço et al., 2018). Clustering-based forecasting refers to the application of un-
supervised machine learning in forecasting tasks. The scope is to increase the performance by employing
the information of data structure and of the existing similarities among the data entries (Goia et al., 2010).
Clustering is a proven method in pattern recognition and data science for deriving the level of similarity
of data points within a set. The outputs of a clustering algorithm are the centroids of the clusters and the
cluster labels, i.e., integer numbers that denote the number of cluster that a specific data entry belongs to
(Xu and Wunsch, 2005).
here are two approaches in clustering-based forecasting: (i) Combination of clustering and supervised
machine learning, and (ii) solely application of clustering. In the first case, a clustering algorithm is used
to split the training set into smaller sub-training sets. These sets contains patterns with high similarity.
Then for each cluster a dedicated forecaster is applied (Chaouch, 2014; Fan et al., 2008). Thus, the number
of forecasting algorithms is equal to the number of clusters. This approach enables to train forecasters
with more similar patterns and eventually achieve better training process. The forecasting systems that
involve clustering are reported to result in lower errors (Fan et al., 2006; Mori and Yuihara, 2001). In the
second case, a clustering algorithm is used to both cluster the load data set and perform the forecasting
(López et al., 2012). In the sole clustering applications, either the centroids of the clusters can be utilised
or the labels. Pattern sequence-based forecasting is an approach that employs the cluster labels. In this
approach, a clustering of all days prior to the test day is held and this results in sequences of labels of
a certain length. Next, the similarity of the predicted day sequence with the historical data sequences is
examined. The load curve of the predicted day is the average of the curves of the days following the same
sequences (Martinez Alvarez et al., 2011).
There is variety of clustering algorithms that have been proposed in forecasting such as the k-means,
fuzzy C-means (FCM), self-organising map (SOM), deterministic annealing (DA), ant colony clustering
ACC, and others. Apart from the clustering effectiveness, a selection criterion for an algorithm is the com-
plexity. k-means and FCM are less complex compared to the SOM that needs a large number of variables
to be calibrated prior to its application. Meta-heuristics algorithms, such as ACC, strongly depend on the
initialisation conditions and the swarm size. Therefore, a comparison of clustering algorithms should take
place to define the most suitable one for the problem under study (Mori and Yuihara, 2001; Li et al., 2008;
Elangasinghe et al., 2014; Wang et al., 2015a).
The assessment of clustering-based forecasting is held via common evaluation metrics for forecasting
tasks. The optimal number of clusters, which is a crucial parameter of a clustering application, is selected
via trial-and-error, i.e., the optimal number corresponds to the lowest forecasting error (Nagi et al., 2011).

2.6.12. Hybrid methods58

Hybrid approaches combine two or more of the above-mentioned advanced methods. In general, when
methods based on AI-based techniques, physical, and statistical approaches are combined together, the re-
sult is often improved forecasting accuracy as a benefit from the inherent integration of the single methods.
The idea is to mix diverse methods with unique features to address the limitations of individual techniques,

58 This subsection was written by Sonia Leva.

56
thus enhancing the forecast performance (Nespoli et al., 2019; Mandal et al., 2012). The performance of
the hybrid methods depends on the performance of the single methods, and these single methods should
be specifically selected for the problem that has to be addressed.
Hybrid methods can be categorised based on the constituent methods, but also considering that these
base methods may not necessarily act only on the forecasting stage but also on data treatment and param-
eters identification stages. In data pre-processing combined approaches, different methods can be used
for decomposing the time series into subseries (Son et al., 2019) or the signal into different frequencies
(Zang et al., 2018), and for classifying the historical data (Huang et al., 2015a). An advantage of such hy-
brid methods is robustness against sudden changes in the values of the main parameters. However, they
require additional knowledge and understanding of the base methods, and have the disadvantage of slow
response time to new data.
The purpose of the parameter selection stage is to optimise the parameters of the model, in terms of
extracting nonlinear features and invariant structures (Behera et al., 2018; Ogliari et al., 2018) but also
in terms of estimation of the parameter adopted for the prediction; for example, meteorological factors
such as temperature, humidity, precipitation, snowfall, cloud, sunshine, wind speed, and wind direction
(Qu et al., 2016). Hybrid methods feature straightforward determination of the parameters with relatively
basic structures. However, the implementation is sometimes challenging, and depends on the knowledge
and expertise of the designer.
Finally, the data post-processing hybrid approaches forecast the residual errors resulted from the fore-
casting model. Since these hybrid methods consider residual errors from the model, they aim in further
improving the predictions of the base methods by applying corrections in the forecasts. However, a dis-
advantage of these hybrid methods is the increased calculation time, as the residual errors must also be
estimated. Also, such hybrid methods are not general and will depend on the field of application.
In many cases, hybrids approaches outperform other (single) approaches such as kNN, NN, and ARIMA-
based models (Mellit et al., 2020). A great example is the hybrid method by (Smyl, 2020), which achieved
the best performance in the M4 forecasting competition. In particular, in energy applications, a combi-
nation of physical and AI-based techniques can lead to improved forecasting performance. Furthermore,
machine learning methods based on historical data of meteorological variables combined with an opti-
mal learning algorithm and weather classification can further improve the forecasting accuracy of single
methods. However, in general, the weak point of such hybrid approaches is that they underperform when
meteorological conditions are unstable (Chicco et al., 2015).

2.7. Methods for intermittent demands and count data

2.7.1. Parametric methods for intermittent demand forecasting59
Demand forecasting is the basis for most planning and control activities in any organisation. Demand
will typically be accumulated in some pre-defined ‘time buckets’ (periods), such as a day, a week or a
month. On many occasions, demand may be observed in every time period, resulting in what is some-
times referred to as ‘non-intermittent demand’. Alternatively, demand may appear sporadically, with no
demand at all in some periods, leading to an intermittent appearance of demand occurrences. Intermittent

59 This subsection was written by Aris A. Syntetos.

57
demand items monopolise the stock bases in the after sales industry and are prevalent in many other in-
dustries, including the automotive, IT, and electronics sectors. Their inventory implications are dramatic
and forecasting their requirements is a very challenging task.
Methods to forecast intermittent demand may broadly be classified as parametric and non-parametric.
The former suppose that future demand can be well represented by a statistical distribution (say Poisson or
Negative Binomial) which has parameters that are unknown but may be forecasted using past data. These
methods are discussed in this sub-section. In the latter, the data are not assumed to follow any standard
probability distribution. Instead, direct methods are used to assess the distributions required for inventory
management. Such methods are discussed in the following sub-section.
Simple Exponential Smoothing (SES) is often used in practice to forecast intermittent demand series.
However, SES fails to recognise that intermittent demand is built from two constituent elements: (i) the
inter-demand intervals, which relate to the probability of demand occurring, and (ii) the demand sizes,
when demand occurs. The former indicates the degree of intermittence, whereas the latter relates to the
behaviour of the positive demands. Croston (1972) showed that this inherent limitation leads to SES be-
ing (positively) biased after a demand occurring period; this is sometimes referred to as an ‘issue point’
bias. Subsequently, he proposed a method that forecasts separately the sizes of demand, when demand
occurs, and the inter-demand intervals. Both forecasts are produced using SES, and the ratio of the former
over the latter gives a forecast of the mean demand per period. Croston’s method was shown by Syntetos
and Boylan (2001) to suffer from another type of bias (inversion bias) and the same researchers (Syntetos
and Boylan, 2005) proposed a modification to his method that leads to approximately unbiased estimates.
This method is known in the literature as the Syntetos-Boylan Approximation (SBA). It has been found
repeatedly to account for considerable empirical inventory forecasting improvements (Eaves and Kings-
man, 2004; Gutierrez et al., 2008; van Wingerden et al., 2014; Nikolopoulos et al., 2016) and, at the time
of writing, it constitutes the benchmark against which other (new) proposed methodologies in the area of
intermittent demand forecasting are assessed.
Croston’s method is based upon the assumption of a Bernoulli demand arrival process. Alternatively,
demand may be assumed to arrive according to a Poisson process. It is also possible to adapt Croston’s
method so that sizes and intervals are updated based on a simple moving average (SMA) procedure instead
of SES. Boylan and Syntetos (2003), Shale et al. (2006), and Syntetos et al. (2015a) presented correction
factors to overcome the bias associated with Croston’s approach under a Poisson demand arrival process
and/or estimation of demand sizes and intervals using an SMA.
For a detailed review of developments in intermittent demand forecasting interested readers are re-
ferred to Boylan and Syntetos (2021).

2.7.2. Non-parametric intermittent demand methods60

In contrast to parametric forecasting, where a demand distribution is assumed and point forecasts
of the unknown distribution parameters are estimated, in the non-parametric forecasting approach, the
demand is not assumed to follow any standard probability distribution. This approach leads directly to
an estimation of the empirical distribution function based on the past data. Two main non-parametric

60 This subsection was written by Mohamed Zied Babai.

58
forecasting approaches have dominated the intermittent demand literature: the bootstrapping approach
and the Overlapping/Non-Overlapping aggregation Blocks approach (Boylan and Syntetos, 2021).
The bootstrapping approach relies upon a resampling (with or without replacement) of the historical
demand data to build the empirical distribution of the demand over a specified interval. This approach
was initially introduced by Efron (1979). Since then, it has been developed by Willemain et al. (2004) and
Zhou and Viswanathan (2011) to deal with intermittent demand items (Babai et al., 2020). Willemain et al.
(2004) have proposed a method that resamples demand data by using a Markov chain to switch between
no demand and demand periods. The empirical outperformance of this method has been shown when
compared to Simple Exponential Smoothing (SES) and Croston’s method. However, the findings of Wille-
main et al. (2004)’s work have been challenged by Gardner and Koehler (2005) and some limitations have
been addressed by (Syntetos et al., 2015b). Zhou and Viswanathan (2011) have developed an alternative
bootstrapping method. Their method samples separately demand intervals and demand sizes and it has
been shown to be associated with a good performance for long lead-times. Teunter and Duncan (2009)
and Hasni et al. (2019a) have developed adjustments of the bootstrapping methods, where the lead-time
demand forecast is adjusted by assuming that the first period in the lead-time bucket corresponds to a
non-zero demand. They have demonstrated the outperformance of the adjusted bootstrapping methods
in a periodic order-up-to-level inventory control system. A review of the bootstrapping methods in the
context of intermittent demand is provided by Hasni et al. (2019b).
Porras and Dekker (2008) were the first to consider aggregation with overlapping and non-overlapping
blocks (OB and NOB) approach in forecasting the demand of spare parts. In the NOB approach, a demand
series is divided into consecutive non-overlapping blocks of time, whereas in OB, at each period the oldest
observation is dropped and the newest is included (Rostami-Tabar et al., 2013). Boylan and Babai (2016)
have compared the statistical and inventory performance of the OB and NOB methods. They found that,
unless the demand history is short, there is a clear advantage of using OB instead of NOB. More recently,
based on extreme value theory (EVT), Zhu et al. (2017) have proposed an improvement of the OB method
that models better the tail of lead-time demand. They have shown that the empirical-EVT method leads to
higher achieved target cycle service levels when compared to the original method proposed by Porras and
Dekker (2008).

2.7.3. Classification methods61

In many application areas, forecasts are required across a wide collection of products, services or lo-
cations. In this situation, it is convenient to introduce classification rules that allow subsets of time series
to be forecasted using the same approaches and methods. Categorisation rules, such as the ABC inven-
tory classification, serve the forecasting function only coincidentally. They do not necessarily align to the
selection of the best forecasting method.
Bartezzaghi et al. (1999) identified five factors that contribute towards intermittence and ‘lumpiness’
(intermittence with highly variable demand sizes): number of potential customers, frequency of customers’
requests, heterogeneity of customers, variety of individual customer’s requests, and correlations between
customers’ requests. These may contribute towards useful classifications, for example by the number
of customers for an item. When this number is low and some of the customers are large, then direct

61 This subsection was written by John E. Boylan.

59
communication with the customers can inform judgemental forecasts. Similarly, if a customer’s requests
are highly variable, then ‘advance demand information’ from customers can help to improve judgemental
estimates. These strategies can be very useful in a business-to-business environment, where such strategies
are feasible.
Within certain modelling frameworks, classification of time series is well established. For example,
within an ARIMA framework (Box et al., 2008), or within a state-space framework for exponential smooth-
ing (Hyndman et al., 2002), series may be classified, for example based on the AIC (Akaike, 1973). It is
more challenging to classify series according to their recommended forecasting method if some of the can-
didate methods, such as Croston’s method, lack a fully satisfactory model base. In the field of intermittent
demand forecasting, Syntetos et al. (2005) proposed the SBC classification scheme, enabling time series to
be classified according to their length of average demand interval and coefficient of variation of demand
sizes (when demand occurs). These rules were based on assumptions of independent and identically dis-
tributed (iid) demand, and a comparison of expected mean square error between methods. The scheme
has been extended by Kostenko and Hyndman (2006) and by Petropoulos and Kourentzes (2015). In an
empirical case-study, Boylan et al. (2008) examined series not necessarily conforming to iid assumptions
and found the rules to be robust to inexact specification of cut-off values. Moon et al. (2013) used logistic
regression to classify time series of demand for spare parts in the South Korean Navy. The classification
was designed to identify superior performance (accuracy and inventory costs) of direct and hierarchical
forecasting methods, based on the serial correlation of demands, the coefficient of variation of demand
volume of spare parts, and the functionality of the naval equipment.
An alternative approach to classification is combination of forecasts. Petropoulos and Kourentzes
(2015) investigated combining standard forecasting methods for intermittent demand (e.g., SES, Croston,
Syntetos-Boylan Approximation). They did not find this to improve accuracy directly, but obtained good
results from the use of combinations of forecasts at different temporal frequencies, using methods selected
from the extended SBC classification scheme.

2.8. Reasoning and mining

2.8.1. Fuzzy logic62
The “classical” Boolean logic is not able to handle for uncertainties and/or vagueness that are necessary
when dealing with many real world problems. This is in part due to the fact that the Boolean logic is based
on only two values (i.e., a statement can only be true or false). Fuzzy logic tries to overcome this issue by
admitting that a statement/variable could be partially true or partially false. Mathematically, the fuzzy
logic framework is based on the work of Zadeh (1965) who introduced the theory of fuzzy sets. The main
point of this theory is the definition of two kinds of sets:

1. Crisp sets are the “classical” sets in the Boolean logic. An element can belong (or not) to a certain set.
2. Fuzzy sets, where an element can belong to the sets with a certain membership grade, with a value that
varies in the interval [0, 1].

The definition of the fuzzy sets allows the framework to take into account the uncertainty and vague-
ness of information. An extension of this approach is related to the fact that a certain variable can assume

62 This subsection was written by Claudio Carnevale.

60
a crisp value (classical theory) or can belong to different fuzzy sets with different membership grade. For
example, in a system implemented to forecast the daily pollutant concentration in atmosphere, one of the
inputs could relate to the weather conditions, such as the wind speed. In the classical approach, the system
must have as an input the value of the wind speed at a certain day. In the fuzzy approach, the input of the
system could be the membership grade of the input variable to three different fuzzy sets: (i) “Not windy”,
(ii) “average windy”, and (iii) “strong windy”. On the other hand, the user of the forecasting system may
be only interested in a classification of the output variable instead of the crisp value. In this case, the fuzzy
approach is applied to the pollutant concentration which could belong with a certain degree to the fuzzy
sets (i) “not polluted day”, (ii) “medium polluted day”, (iii) “high polluted day”, and (iv) “critical polluted
day”.
In fuzzy theory, each fuzzy set is characterised by a (generally nonlinear) function, called the member-
ship function, linking crisp values to the membership of the different sets. The association of a crisp value
to its membership for a set is called fuzzyfication, while the inverse operation (from a membership value
to a crisp value) is called defuzzification. As with the logic theory, the inference system assumes a key role
in the fuzzy theory. A Fuzzy Inference System (FIS) allows the interpretation of the membership grades of
the input variable(s) and, given some sets of fuzzy rules, assigns the corresponding values to the output
variable(s). In the literature, two main fuzzy inference systems are presented:

1. Mamdani system (Mamdani and Assilian, 1975), where both the input and output of the inference
system are membership functions.
2. Sugeno system (Sugeno, 1985), where the output of the inference system is a crisp value, usually
obtained by applying a linear function to the defuzzified value of the input.

2.8.2. Association rule mining63

Association rule mining is an exploratory data-driven approach which is able to automatically and
exhaustively extract all existing correlations in a data set of categorical features. It is a powerful but com-
putationally intensive technique, successfully applied in different forecasting contexts (Acquaviva et al.,
2015; Apiletti and Pastor, 2020; Di Corso et al., 2018). Its results are in a human-readable form.
The data set must be in the form of transactions, i.e., a collection of events, each described by categorical
features. If the phenomena under analysis are modelled by continuous-valued variables, discretisation can
be applied to obtain a suitable data set.
Association rule mining core task is the frequent itemset extraction, which consists in finding frequently-
occurring relationships among items in a data set (Han et al., 2011). Given a data set of records charac-
terised by several attributes, an item refers to a pair of (attribute=value), while a set of items is called
itemset. The support count of an itemset is the number of records r containing that itemset. The support
of an itemset is the percentage of records containing it with respect to the total number of records in the
data set. An itemset is frequent when its support is greater than or equal to a minimum support threshold.
An association rule is an implication in the form A → B, where A and B are disjoint itemsets (i.e.,
A ∩ B = ∅) (Tan et al., 2005). A is called rule body or antecedent and B rule head or consequent.
To evaluate the quality of an association rule, the support, confidence, and lift metrics are commonly
exploited (Han et al., 2011). Rule support is the fraction of records containing both A and B, indicating the

63 This subsection was written by Daniele Apiletti.

61
probability that a record contains every item in these itemsets. The support of the rule is computed as the
support of the union of A and B.
Rule confidence represents the strength of the implication, and is the conditional probability that a
transaction containing A also contains B, P (B|A), i.e., the proportion of records that contain A with respect
to those that also contain B.
Finally, the lift of a rule measures the correlation between antecedent and consequent. It is defined as
the ratio between the rule A → B confidence and the support of B. A lift ratio equal to 1.0 implies that
itemsets A and B are not correlated. A lift higher than 1.0 indicates a positive correlation, meaning that
the occurrence of A likely leads to the occurrence of B with the given confidence. The greater the lift, the
stronger the association. Finally, a lift lower than 1.0 indicates a negative correlation between A and B.
The problem of association rule mining consists in the extraction of all the association rules having
rule support and confidence greater than the respective support and confidence thresholds, MinConf and
MinSup, defined as parameters of the mining process (Tan et al., 2005). These thresholds allow to control
the statistical relevance of the extracted rules.
The process of rule mining consists of two steps. The first step is the computation of frequent itemsets,
i.e., itemsets with support greater or equal to MinSup. The second step is the extraction of association
rules from frequent itemsets. Let be F a frequent itemset, hence having a support higher than MinSup,
pairs A and B = F − A are derived so that the confidence of A → B is higher than MinConf . The first step of
the process is the most computationally expensive. Thus, several algorithms have been proposed to solve
the problem of frequent itemset extraction (Zaki, 2000), some specifically addressing high-dimensionality
issues (Apiletti et al., 2017, 2015). Despite being computationally demanding, association rule mining is
an exhaustive approach, i.e., all and only statistically relevant correlations are extracted.

2.8.3. Forecasting with text information64

Text data, such as social media posts, scholar articles and company reports, often contains valuable
information that can be used as predictors in forecasting models (Aggarwal and Zhai, 2012). Before ex-
tracting useful features from the data, a text document needs to be cleaned and normalised for further
processing. The first step of preparing the data is to filter out stop words – the words that do not add much
meaning to a sentence, e.g., “a”, “is”, and “me”. For grammatical reasons, it is necessary for documents
to use different forms of words. Stemming and lemmatisation can be applied to reduce inflectional forms
or relate different forms of a word to a common base form. Stemming often chops off the end of a word,
while lemmatisation uses vocabularies and morphological analysis to return the base form (or the lemma)
of a word (Lovins, JB, 1968; Manning et al., 2008). For example, the word “industries” will be turned into
“industri” or “industry” if stemming or lemmatisation is applied.
To model and analyse text data, we need to transform it into numerical representations so the fore-
casting models can process them as predictors. One way of transforming the data is through sentiment
analysis. Sentiment analysis is often applied to detect polarity within customer materials such as reviews
and social media posts (Das and Chen, 2007; Archak et al., 2011). An easy way to obtain the sentiment
score of a word is to look it up in a happiness dictionary (for example, the hedonometer dictionary, Hedo-
nometer, 2020). Another common way of representing the sentiment is to use a vector of numeric values

64 This subsection was written by Xiaojia Guo.

62
that denote the word’s positivity, negativity and neutrality based on existing lexical databases such as the
WordNet (Godbole et al., 2007; Baccianella et al., 2010). Once the sentiment of each word is calculated, we
can apply an aggregation algorithm (e.g., simple average) to measure the sentiment of an entire sentence
or paragraph.
In scholar articles and company reports, context features might be more important than sentiments.
The bag-of-words model and word embeddings are often applied to generate numeric representations of
such text. A bag-of-words model simply returns a matrix that describes the occurrence of words within a
document (Goldberg, 2017). When we use this matrix as input to a forecasting model, each word count can
be considered as a feature. The Word2Vec method is a widely used embedding method that is built based
on the context of a word. Specifically, it trains a two-layer neural network that takes each word as an input
to predict its surrounding words. The weights from the input layer to the hidden layer are then utilised
as the numerical representation for the input word (Le and Mikolov, 2014). Once the text is turned into
numeric representations, they can be used as predictors in any forecasting models. The most challenging
part in this process is to find the right technique to extract features from the text.
In terms of software implementation, the Natural Language Toolkit (NLTK) and SpaCy library in
Python can be applied to remove stop words and stem or lemmatise text (Loper and Bird, 2002; Hon-
nibal, 2015). The bag-of-words technique is also available in NLTK. A particular implementation of the
Word2Vec model is available on Google code (2013). Moreover, a public data set of movie reviews that is
commonly studied in literature is available from the Stanford NLP Group (2013).

2.9. Forecasting by aggregation

2.9.1. Cross-sectional hierarchical forecasting65
In many applications time series can be aggregated at several levels of aggregation based on geographic
or logical reasons to form hierarchical structures. These are called hierarchical time series. On a retail
business for example, individual sales of products at the bottom-level of the hierarchy can be grouped in
categories and families of related products at increasing aggregation levels, with the total sales of the shop
or distribution centre at the top level (Pennings and van Dalen, 2017; Oliveira and Ramos, 2019; Villegas
and Pedregal, 2018). The forecasts of hierarchical time series produced independently of the hierarchical
structure generally will not add up according to the aggregation constrains of the hierarchy, i.e., they
are not coherent. Therefore, hierarchical forecasting methods that generate coherent forecasts should be
considered to allow appropriate decision-making at the different levels. Actually, by taking advantage
of the relationships between the series across all levels these methods have shown to improve forecast
accuracy (Athanasopoulos et al., 2009; Shang and Hyndman, 2017; Yagli et al., 2019).
The most common approaches to hierarchical forecasting are bottom-up and top-down. In the bottom-
up approach forecasts for each time series at the bottom-level are first produced and then these are added
up to obtain forecasts for all other series at the hierarchy (Dunn et al., 1976). Since forecasts are obtained
at the bottom-level no information is lost due to aggregation. In the top-down approach forecasts for the
top-level series are first generated and then these are disaggregated generally using historical proportions
to obtain forecasts for the bottom-level series, which are then aggregated (Gross and Sohl, 1990). Hyndman
et al. (2011) showed that this approach introduces bias to the forecasts.

65 This subsection was written by Patrı́cia Ramos.

63
 Recent research on hierarchical forecasting tackles the problem using a two-stage approach. Forecasts
for the series at all levels of the hierarchy are first obtained independently without considering any aggre-
gation constrains (we refer to these as base forecasts). Then, base forecasts are adjusted so that they become
coherent (we refer to these as reconciled forecasts). This adjustment is achieved by a matrix that maps the
base forecasts into new bottom-level forecasts which are then added up (Hyndman et al., 2011).
Wickramasuriya et al. (2019) found the optimal solution for this matrix, which minimises the trace of
the covariance matrix of the reconciled forecast errors (hence MinT reconciliation). This optimal solution
is based on the covariance matrix of the base forecast errors which incorporates the correlation structure
of the hierarchy. Wickramasuriya et al. (2019) presented several alternative estimators for this covariance
matrix: (i) proportional to the identity which is optimal only when base forecast errors are uncorrelated
and equivariant (referred to as OLS), (ii) proportional to the sample covariance estimator of the in-sample
one-step-ahead base forecast errors with off-diagonal elements null accounts for the differences in scale
between the levels of the hierarchy (referred to as WLS), (iii) proportional to the previous estimator unre-
stricted also accounts for the relationships between the series (referred to as MinT-Sample), and (iv) pro-
portional to a shrinkage estimator based on the two previous estimators, parameterising the shrinkage in
terms of variances and correlations, accounts for the correlation across levels (referred as MinT-Shrink).
More recently these techniques were extended to probabilistic forecasting (Ben Taieb et al., 2020).
When base forecasts are probabilistic forecasts characterised by elliptical distributions, Panagiotelis et al.
(2020) showed that reconciled probabilistic forecasts also elliptical can be obtained analytically. When it
is not reasonable to assume elliptical distributions, a non-parametric approach based on bootstrapping
in-sample errors can be used.

2.9.2. Temporal aggregation66

Temporal aggregation is the transformation of a time series from one frequency to another of lower
frequency. As an example, a time series of length n that is originally sampled at a monthly frequency can
be transformed to a quarterly series of length n/3 by using equally-sized time buckets of three periods each.
It is usually applied in an non-overlapping manner, but overlapping aggregation can also be considered.
The latter is preferred in the case when the original series is short, but has the disadvantage of applying
lower weights on the few first and last observations of the series and introducing autocorrelations (Boylan
and Babai, 2016).
Temporal aggregation is appealing as it allows to investigate the original series through different lenses.
By changing the original frequency of the data, the apparent series characteristics also change. In the case
of slow-moving series, temporal aggregation leads to decrease of intermittence (Nikolopoulos et al., 2011).
In the case of fast-moving series, higher levels of aggregation (i.e., lower frequencies) allow for better
modelling of trend patterns, while lower aggregation levels (i.e., higher frequencies) are more suitable for
capturing seasonal patterns (Spithourakis et al., 2014; Kourentzes et al., 2014).
Research has found evidence of improved forecasting performance with temporal aggregation for both
slow (Nikolopoulos et al., 2011) and fast (Spithourakis et al., 2011) moving time series. This led to charac-
terise temporal aggregation as a “self-improving mechanism”. The good performance of temporal aggrega-
tion was reconfirmed by Babai et al. (2012), who focused on its utility performance rather than the forecast

66 This subsection was written by Fotios Petropoulos.

64
error. However, one challenge with single levels of aggregation is the choice of a suitable aggregation level
for each series (Kourentzes et al., 2017).
Instead of focusing on a single aggregation level, Andrawis et al. (2011), Kourentzes et al. (2014),
Petropoulos and Kourentzes (2014), and Petropoulos and Kourentzes (2015) suggested the use of mul-
tiple levels of aggregation, usually abbreviated as MTA (multiple temporal aggregation). This not only
tackles the need to select a single aggregation level, but also partly addresses the issue of model uncer-
tainty, instead of relying on model selection and parametrisation at a single aggregation level. Using this
property, Kourentzes et al. (2017) showed that MTA will typically lead to more accurate forecasts, even if
in theory suboptimal. Different frequencies allow for better identification of different series patterns, so
it is intuitive to consider multiple temporal levels and benefit from the subsequent forecast combination
across frequencies. Kourentzes and Petropoulos (2016) showed how multiple temporal aggregation can be
extended to incorporate exogenous variables. However, forecasting at a single level of aggregation can still
result in better performance when the seasonal pattern is strong Spiliotis et al. (2019b, 2020c).
Athanasopoulos et al. (2017) expressed multiple temporal aggregation within the hierarchical forecast-
ing framework using the term “temporal hierarchies”. Temporal hierarchies allow for the application of
established hierarchical reconciliation approaches directly to the temporal dimension. Jeon et al. (2019)
show how temporal hierarchies can be used to obtain reconciled probabilistic forecasts, while Spiliotis et al.
(2019b) explored empirical bias-adjustment strategies and a strategy to avoid excessive seasonal shrinkage.
Nystrup et al. (2020) proposed estimators for temporal hierarchies suitable to account for autocorrelation
in the data. Finally, Kourentzes and Athanasopoulos (2020) applied temporal hierarchies on intermittent
data, and showed that higher aggregation levels may offer structural information which can improve the
quality of the forecasts.

2.9.3. Cross-temporal hierarchies67

In the last two subsections (2.9.1 and 2.9.2), we saw two complimentary hierarchical structures, cross-
sectional and temporal. Although the machinery behind both approaches is similar, often relying on the
hierarchical framework by Hyndman et al. (2011) and Athanasopoulos et al. (2009), and work that followed
from these (particularly Wickramasuriya et al., 2019), they address different forecasting problems. Cross-
sectional hierarchies change the unit of analysis but are fixed in the period of analysis. For example, a
manufacturer may operate using a hierarchy across products. The different nodes in the hierarchy will
correspond to different products, product groups, super-groups, and so on, but will all refer to the same
period, for example, a specific week. Temporal hierarchies do the opposite, where the unit of analysis is
fixed, but the period is not. For example, we may look at the sales of a specific Stock Keeping Unit (SKU) at
a daily, weekly, monthly, quarterly and annual levels. However, one can argue that having annual forecasts
at the SKU level may not be useful. Similarly, having aggregate sales across an organisation at a weekly
frequency is also of little value.
In connecting these to organisational decisions, we can observe that there is only a minority of problems
that either cross-sectional or temporal hierarchies are natural, as typically decisions can differ across both
the unit and the period (planning horizon) of analysis. In the latter case, both hierarchical approaches

67 This subsection was written by Nikolaos Kourentzes.

65
are more akin to statistical devices that can improve forecast accuracy through the use of forecast com-
binations, rather than satisfy the motivating argument behind hierarchical forecasting that is to provide
coherent predictions for decisions at different levels of the hierarchy.
Cross-temporal hierarchies attempt to overcome this limitation, providing coherent forecasts across
all units and periods of analysis, and therefore a common outlook for the future across decision-makers
at different functions and levels within an organisation. The literature remains sparse on how to con-
struct cross-temporal forecasts, as the size of the hierarchy can easily become problematic. Kourentzes
and Athanasopoulos (2019) propose a heuristic approach to overcome the ensuing estimation issues. The
approach works by compartmentalising the estimation. First, they obtain estimates of the cross-sectional
reconciliation weights for each temporal level of the hierarchy. Then, these are combined across temporal
levels, to a unique set that satisfies all coherency constraints. Using these combined weights, they obtain
the reconciled bottom level forecasts, which can be aggregated as needed. Although they recognise that
their approach can result in suboptimal results in terms of reconciliation errors, it guarantees coherent
forecasts. Cross-temporal forecasts are more accurate than either temporal or cross-sectional hierarchical
forecasts and provide a holistic view of the future across all planning levels and demarcations. Spiliotis
et al. (2020c) also identify the problem, however, they do not focus on the coherency of forecasts and pro-
pose a sequential reconciliation across the two dimensions. This is shown to again be beneficial, but it
does not achieve coherency. Arguably one can adapt the iterative correction algorithm by Kourentzes and
Athanasopoulos (2020) to enforce coherency in this approach as well.

2.9.4. Ecological inference forecasting68

Ecological inference forecasting (EIF) aims to predict the inner-cells values of a set of contingency ta-
bles when only the margins are known. It defines a fundamental problem in disciplines such as political
science, sociology and epidemiology (Salway and Wakefield, 2004). Cleave et al. (1995), Greiner (2007)
and Pavia et al. (2009) describe other areas of application. The fundamental difficulty of EIF lies in the fact
that this is a problem with more unknowns than observations, giving rise to concerns over identifiability
and indeterminacy: many sets of substantively different internal cell counts are consistent with a given
marginal table. To overcome this issue, a similarity hypothesis (and, sometimes, the use of covariates) is
routinely assumed. The basic hypothesis considers that either conditional row (underlying) probabilities
or fractions are similar (related) among contingency tables (Greiner and Quinn, 2010). The covariations
among row and column margins of the different tables are then used to learn about the internal cells.
The above hypothesis is not a cure-all to the main drawback of this approach. EIF is exposed to the so-
called ecological fallacy (Robinson, 1950): the presence of inconsistencies in correlations and association
measures across different levels of aggregation. This is closely related to the well-known Simpson’s Paradox
(Simpson, 1951). In this setting, the ecological fallacy is manifested through aggregation bias (Wakefield,
2004) due to contextual effects and/or spatial autocorrelation (Achen and Phillips Shively, 1995). This
has led many authors to disqualify ecological inference forecasts (see, for example, Freedman et al., 1998;
Tam Cho, 1998; Anselin and Tam Cho, 2002; Herron and Shotts, 2004) and many others to study under
which circumstances ecological inference predictions would be reliable (Firebaugh, 1978; Gelman et al.,

68 This subsection was written by Jose M. Pavı́a.

66
2001; Forcina and Pellegrino, 2019). Despite the criticisms, many algorithms for solving the EIF prob-
lem can be found in the literature, mainly from the ecological regression and mathematical programming
frameworks (some of them available in functions of the R statistical software).
The ecological regression literature has been prolific since the seminal papers of Goodman (1953, 1959)
and Duncan and Davis (1953) and is undergoing a renaissance after King (1997): new methods easily
generalisable from 2 × 2 tables to R × C tables have been proposed (King et al., 1999; Rosen et al., 2001),
the geographical dimension of the data is being explicitly considered (Calvo and Escolar, 2003; Puig and
Ginebra, 2015), and new procedures combining aggregated and individual level data, including exit polls,
are introduced (Glynn and Wakefield, 2010; Greiner and Quinn, 2010; Klima et al., 2019). See King et al.
(2004) for a wide survey and Klima et al. (2016) and Plescia and De Sio (2018) for an extensive evaluation
of procedures. In mathematical programming exact and inequality constraints for the inner-cell values
are incorporated in a natural way. Hence, this approach has shown itself to be a proper framework for
generating ecological inference forecasts. The proposals from this approach can be traced back to Hawkes
(1969) and Irwin and Meeter (1969). After them, some key references include McCarthy and Ryan (1977),
Tziafetas (1986), Corominas et al. (2015), and Romero et al. (2020). Solutions based on other strategies,
for instance, entropy maximization, have been also suggested (see, for example, Johnston and Pattie, 2000;
Bernardini Papalia and Fernandez Vazquez, 2020).

2.10. Forecasting with judgment

2.10.1. Judgmental forecasting69
People may use judgment alone to make forecasts or they may use it in combination with statistical
methods. Here the focus is on pure judgmental forecasting. Different types of judgment heuristic (mental
‘rules of thumb’) can be used to make forecasts. The heuristic used depends on the nature of the informa-
tion available to the forecaster (Harvey, 2007).
Consider cases where the only relevant information is held in the forecaster’s memory. For example,
someone might be asked whether Manchester United or Burnley will win next week’s match. Here one
memory-based heuristic that might be applicable is the recognition heuristic: if one recognises one object
but not the other, then one should infer that the recognised object has higher value (Goldstein and Gigeren-
zer, 2002). In the above example, most people who recognise just one of the teams would be likely to make
a correct forecast that Manchester United will win (Ayton et al., 2011). The availability heuristic is another
memory-based heuristic that may be applicable: objects that are most easily brought to mind are those
which are more likely. Thus, if we are asked which team is likely to come top of the premier league, we
would say Manchester United if that is the one that most easily comes to mind. The availability heuristic
is often effective because more likely events are encountered more often and more recently and are hence
better remembered. However, it can be disrupted by, for example, greater media coverage of more unlikely
(and hence more interesting) events.
Consider next cases in which forecasters possess information about values of one or more variables
correlated with the variable to be forecast. For example, teachers may wish to forecast the grades of their
students in a final examination on the basis of past records of various other measures. Kahneman and

69 This subsection was written by Nigel Harvey.

67
Tversky (1973) suggested that people use the representativeness heuristic to deal with this type of situa-
tion. Forecasters first select a variable that they think is able to represent the one that must be predicted.
For example, a teacher may consider that frequency in attending voluntary revision classes represents a
student’s ability in the final examination. Thus, if a student attended 15 of the 20 revision classes, they are
likely to obtain 75% in the final examination.
Finally, consider situations in which people forecast future values of a variable on the basis of a record
of previous values of that variable. There is some evidence that, when forecasting from time series, people
use anchor-and-adjustment heuristics (Hogarth and Makridakis, 1981; Lawrence and O’Connor, 1992). For
example, (i) when forecasting from an upward trended series, they anchor on the last data point and then
make an upward adjustment to take the trend into account and (ii) when forecasting from an untrended
series containing autocorrelation, they anchor on the last data point and make an adjustment towards the
mean to take the autocorrelation into account.
Kahneman (2011) and others have divided cognitive processes into those which are intuitive (System
1) and those which are deliberative (System 2). We have discussed only intuitive processes underlying
judgmental forecasting (Gigerenzer, 2007). However, they can be supplemented by deliberative (System
2) processes (Theocharis and Harvey, 2019) in some circumstances.

2.10.2. Judgmental adjustments of computer-based forecasts70

Judgmental adjustments to algorithmic computer-based forecasts can enhance accuracy by incorporat-
ing important extra information into forecasts (McNees, 1990; Fahimnia et al., 2020; Perera et al., 2019).
However, this information is sometimes used inefficiently (Fildes et al., 2019a). Adjustments are also often
made for other reasons and these can be detrimental to accuracy (Fildes et al., 2009; Franses and Legerstee,
2009a).
People may ‘egocentrically discount’ a computer’s forecasts when its rationale is less clear than their
own reasoning (Bonaccio and Dalal, 2006). They can also be less tolerant of errors made by algorithms
than those made by humans (Önkal et al., 2009; Dietvorst et al., 2015). The random errors associated with
algorithmic forecasts, and the salience of rare large errors, can therefore lead to an unjustified loss of trust
in computer forecasts. Adjustments may also give forecasters a sense of ownership of forecasts or be used
to justify their role (Önkal and Gönül, 2005).
Computer-based forecasts are designed to filter randomness from time-series. In contrast, humans tend
to perceive non-existent systematic patterns in random movements (O’Connor et al., 1993; Reimers and
Harvey, 2011) and apply adjustments to reflect them. This can be exacerbated by the narrative fallacy
(Taleb, 2008), where people invent stories to explain these random movements, and hindsight bias (Fis-
chhoff, 2007), where they believe, in retrospect, that these movements were predictable. Recent random
movements, and events, are particularly likely to attract undue attention, so long-run patterns identified
by the computer are given insufficient weight (Bolger and Harvey, 1993). Damaging interventions are also
probable when they result from political interference (Oliva and Watson, 2009) or optimism bias (Fildes
et al., 2009), or when they reflect information already factored into the computer’s forecast, leading to
double counting (Van den Broeke et al., 2019).

70 This subsection was written by Paul Goodwin.

68
 How can interventions be limited to occasions when they are likely to improve accuracy? Requiring
people to document reasons justifying adjustments can reduce gratuitous interventions (Goodwin, 2000b).
Explaining the rationale underlying statistical forecasts also improved adjustment behaviour when series
had a simple underlying pattern in a study by Goodwin and Fildes (1999). However, providing guidance
on when to adjust was ineffective in an experiment conducted by (Goodwin et al., 2011), as was a restriction
preventing people from making small adjustments.
When determining the size of adjustments required, decomposing the judgment into a set of easier tasks
improved accuracy in a study by Webby et al. (2005). Providing a database of past outcomes that occurred
in circumstances analogous to those expected in the forecast period also improved adjustments in a study
by Lee et al. (2007). Outcome feedback, where the forecaster is informed of the most recent outcome is
unlikely to be useful since it contains noise and exacerbates the tendency to over-focus on recent events
(Goodwin and Fildes, 1999; Petropoulos et al., 2016). However, feedback on biases in adjustments over
several recent periods may improve judgments (Petropoulos et al., 2017). Feedback will be less useful
where interventions are rare so there is insufficient data to assess performance.

2.10.3. Judgmental model selection71

Forecasters – practitioners and researchers alike – use Forecast Support Systems (FSS) in order to per-
form their forecasting tasks. Usually, such an FSS allows the forecaster to load their historical data and
they can then apply many different types of forecasting techniques to the selected data. The idea is that
the forecaster selects the method which leads to highest forecast accuracy. Yet, there is no universally ‘best’
method, as it depends on the data that is being forecasted. Thus, selection is important in achieving high
accuracy. But how does this selection occur?
Research on judgmental selection of statistical forecasts is limited in quantity. Lawrence et al. (2002)
found that participants were not very adept at selecting good forecasting algorithms from a range offered
to them by an FSS and had higher error than those who were presented with the optimal algorithm by an
FSS. Petropoulos et al. (2018b) compared judgmental selection of forecasting algorithms with automatic
selection based on predetermined information criteria. They found that judgmental selection was bet-
ter than automatic selection at avoiding the ‘worst’ models, but that automatic selection was superior at
choosing the ‘best’ ones. In the end, overall accuracy of judgmental selection was better than that of algo-
rithmic selection. If their experiment had included more variation of the data (trends, fractals, different
autoregressive factors) and variation of proposed models, this could possibly have led to better algorith-
mic than judgmental performance (Harvey, 2019). Time series that are more complex will place a higher
cognitive load on judgmental selection. This was confirmed in a study by Han et al. (2019), who used
an electroencephalogram (EEG) for the comparison of judgmental forecast selection versus (judgmental)
pattern identification. They found that pattern identification outperformed forecast selection, as the latter
required a higher cognitive load, which in turn led to a lower forecasting accuracy.
It is likely that, in practice, judgmental selection is much more common than automatic selection.
This preference for human judgment over advice from an algorithm has been shown in an experiment by
Önkal et al. (2009). But how apt are forecasters in distinguishing ‘good’ models from ‘bad’ models? This
was investigated by De Baets and Harvey (2020) in an experiment. People were asked to select the best

71 This subsection was written by Shari De Baets.

69
performing model out of a choice of two different qualities (accuracies) of models (different combinations
of good versus medium versus bad). People’s choice outperformed forecasts made by averaging the model
outputs, lending credence to the views of Fifić and Gigerenzer (2014). The performance of the participants
improved with a larger difference in quality between models and a lower level of noise in the data series.
In a second experiment, De Baets and Harvey (2020) found that participants adjusted more towards the
advice of what they perceived to be a good quality model than a medium or bad quality one.
Importantly, in selecting an algorithm and seeing it err, people are quick to abandon it. This phe-
nomenon is known as ‘algorithm aversion’ (Dietvorst et al., 2015) and is due to a ‘perfection schema’ we
have in our heads where algorithms are concerned (Madhavan and Wiegmann, 2007). We do not expect
them to ‘fail’ and thus react strongly when they do. While a model may not perform as it should for a par-
ticular dataset and may thus elicit algorithm aversion for that particular method, one should not abandon
it for all datasets and future forecasts.

2.10.4. Panels of experts72

Panels of experts are often used in practice to produce judgmental forecasts. This is especially true in
cases with limited available quantitative data and with the level of uncertainty being very high. In this
section, three methods for eliciting judgmental forecasts from panels of experts are presented: the Delphi
method, interaction groups (IG), and structured analogies (SA).
The Delphi method is centred around organising and structuring group communication (Rao et al.,
2010), which aims to achieve a convergence of opinion on a specific real-world issue. It is a multiple-
round survey in which experts participate anonymously to provide their forecasts and feedback (Rowe
and Wright, 2001). At the end of each round, the facilitator collects and prepares statistical summaries
of the panel of experts’ forecasts. These summaries are presented as feedback to the group, and may be
used towards revising their forecasts. This loop continues until a consensus is reached, or the experts in
the panel are not willing to revise their forecasts further. In some implementations of the Delphi method,
justification of extreme positions (forecasts) is also part of the (anonymous) feedback process. The Delphi
method results in a more accurate outcome in the decision-making process (Dalkey, 1969; Steurer, 2011).
Rowe and Wright (2001) mentioned that, by adopting the Delphi method, groups of individuals can pro-
duce more accurate forecasts than simply using unstructured methods. A drawback of the Delphi method
is the additional cost associated with the need to run multiple rounds, extending the forecasting process
as well as increasing the potential drop-out rates. On the other hand, the anonymity in the Delphi method
eliminates issues such as groupthink and the ‘dominant personalities’ effects (Van de Ven and Delbeco,
1971).
The IG method suggests that the members of the panel of experts actively interact and debate their
points to the extent they have to reach an agreement on a common forecast (Litsiou et al., 2019). Sniezek
and Henry (1989) found that members of interacting groups provide more accurate judgments compared
to individuals. However, there is mixed evidence about the forecasting potential of IG (Scott Armstrong,
2006; Boje and Murnighan, 1982; Graefe and Armstrong, 2011). Besides, the need for arranging and facil-
itating meetings for the IG makes it a less attractive option.

72 This subsection was written by Konstantia Litsiou.

70
 Another popular approach to judgmental forecasting using panels of experts is SA, which refers to the
recollection of past experiences and the use analogies (Green and Armstrong, 2007). In the SA method, the
facilitator assembles a panel of experts. The experts are asked to recall and provide descriptions, forecasts,
and similarities/differences for cases analogous to the target situation, as well as a similarity ranking for
each of these analogous cases. The facilitator gathers the lists of the analogies provided by the experts,
and prepares summaries, usually using weighted averages of the recalled cases based on their similarity to
the target situation. Semi-structured analogies (sSA) have also been proposed in the literature, where the
experts are asked to provide a final forecasts based on the analogous cases they recalled, which essentially
reduces the load for the facilitator (Nikolopoulos et al., 2015). Nikolopoulos et al. (2015) supported that the
use of SA and IG could result to forecasts that are 50% more accurate compared to unstructured methods
(such as unaided judgment). One common caveat of using panels of experts is the difficulty to identify
who a real expert is. Engaging experts with high level of experience, and encouraging the interaction of
experts are also supported by Armstrong and Green (2018).

2.10.5. Scenarios and judgmental forecasting73

Scenarios provide exhilarating narratives about conceivable futures that are likely to occur. Through
such depictions they broaden the perspectives of decision makers and act as mental stimulants to think
about alternatives. Scenarios enhance information sharing and provide capable tools for communication
within organisations. By virtue of these merits, they have been widely used in corporate planning and
strategy setting since 1960’s (Godet, 1982; Schoemaker, 1991; Wright and Goodwin, 1999, 2009; Goodwin
and Wright, 2010). Even though utilisation of scenarios as decision advice to judgmental forecasting has
been proposed earlier (Schnaars and Topol, 1987; Bunn and Salo, 1993), the research within this domain
remained limited until recently when the interest in the subject has rekindled (Önkal et al., 2013; Goodwin
et al., 2019b; Wicke et al., 2019).
The recent research has used behavioural experimentation to examine various features of scenarios and
their interactions with judgmental adjustments of model-based forecasts. Önkal et al. (2013) explored
the ‘content’ effects of scenarios where through the narration either a bleak/negative future (a pessimistic
scenario) or a bright/positive future (an optimistic scenario) was portrayed. On a demand forecasting con-
text for mobile phones, the participants first received time-series data, model-based forecasts and then
asked to generate point and interval forecasts as well as provide a confidence measure. With respect to
the existence of scenarios, there were four conditions where the participants may receive: (i) no scenarios,
(ii) optimistic scenarios, (iii) pessimistic scenarios, and (iv) both scenarios. Findings indicate that deci-
sion makers respond differently to optimistic and pessimistic scenarios. Receiving optimistic scenarios
resulted in making larger adjustments to the model-based forecasts. At the same time, led to an increased
confidence of the participants in their predictions. On the other hand, participants who received negative
scenarios tend to lower their predictions the most among the four groups. An intriguing finding was the
balancing effect of scenarios on the interval forecast symmetry. The lower interval bounds were adjusted
upwards the most towards the centre-point of the interval (i.e., model-based predictions) when optimistic
scenarios were received. Similarly, the upper bounds were adjusted downwards the most towards the
centre-point of the interval in the presence of pessimistic scenarios.

73 This subsection was written by M. Sinan Gönül.

71
 The prospects of receiving a single scenario versus multiple scenarios were further explored in Good-
win et al. (2019b). The researchers investigated whether assimilation or contrast effects will occur when
decision makers see optimistic (pessimistic) forecasts followed by pessimistic (optimistic) ones compared
against receiving a single scenario in solitude. In case of assimilation, a scenario presenting an opposing
world view with the initial one would cause adjustments in the opposite direction creating an offset effect.
On the other hand, in case of contrast, the forecasts generated after the initial scenarios would be adjusted
to more extremes when an opposing scenario is seen. In two experiments conducted in different contexts
the researchers found resilient evidence for contrast effects taking place. Interestingly, seeing an opposing
scenario also increased the confidence of the forecasters in their initial predictions.
In terms of the effects of scenario presence on the forecasting performance, however, the experimen-
tal evidence indicates the benefits are only circumstantial. Goodwin et al. (2019a) found that providing
scenarios worsened forecast accuracy and shifted the resultant production order decisions further away
from optimality. Despite this performance controversy, the decision makers express their fondness in re-
ceiving scenarios and belief in their merits (Önkal et al., 2013; Goodwin et al., 2019b). Therefore, we need
more tailored research on scenarios and judgmental forecasting to reveal the conditions when scenarios
can provide significant improvements to the forecasting accuracy.

2.10.6. Trusting model and expert forecasts74

Defined as “firm belief in the reliability, truth, and ability of someone/something” (Oxford English
Dictionary), trust entails accepting vulnerability and risk (Rousseau et al., 1998). Given that forecasts are
altered or even discarded when distrusted by users, examining trust is a central theme for both forecasting
theory and practice.
Studies examining individual’s trust in model versus expert forecasts show that individuals often dis-
trust algorithms (Meehl, 2013; Burton et al., 2020) and place higher trust on human advice (Diab et al.,
2011; Eastwood et al., 2012). We live in an era where we are bombarded with news about how algorithms
get it wrong, ranging from COVID-19 forecasts affecting lockdown decisions to algorithmic grade pre-
dictions affecting university admissions. Individuals appear to prefer forecasts from humans over those
from statistical algorithms even when those forecasts are identical (Önkal et al., 2009). Furthermore, they
lose trust in algorithms quicker when they see forecast errors (Dietvorst et al., 2015; Prahl and Van Swol,
2017). Such ‘algorithm aversion’ and error intolerance is reduced when users have opportunity to adjust
the forecasting outcome, irrespective of the extent of modification allowed (Dietvorst et al., 2018). Feed-
back appears to improve trust, with individuals placing higher trust in algorithms if they can understand
them (Seong and Bisantz, 2008). Overuse of technical language may reduce understanding of the fore-
cast/advice, in turn affecting perceptions of expertise and trustworthiness (Joiner et al., 2002). Explana-
tions can be helpful (Goodwin et al., 2013b), with their packaging affecting judgments of trustworthiness
(Elsbach and Elofson, 2000). Algorithmic appreciation appears to easily fade with forecasting expertise
(Logg et al., 2019), emphasising the importance of debiasing against overconfidence and anchoring on
one’s own predictions.
Trusting experts also presents challenges (Hendriks et al., 2015; Hertzum, 2014; Maister et al., 2012).
Expert forecasts are typically seen as predisposed to group-based preconceptions (Brennan, 2020; Vermue

74 This subsection was written by Dilek Önkal.

72
et al., 2018), along with contextual and motivational biases (Burgman, 2016). Misinformed expectations,
distorted exposures to ‘forecast failures’, and over-reliance on one’s own judgments may all contribute to
distrusting experts as well as algorithms.
Credibility of forecast source is an important determinant in gaining trust (Önkal et al., 2019). Studies
show that the perceived credibility of system forecasts affects expert forecasters’ behaviours and trust
(Alvarado-Valencia and Barrero, 2014), while providing information on limitations of such algorithmic
forecasts may reduce biases (Alvarado-Valencia et al., 2017). Previous experience with the source appears
to be key to assessing credibility (Hertzum, 2002) and trust (Cross and Sproull, 2004). Such ‘experienced’
credibility appears to be more influential on users’ acceptance of given forecasts as opposed to ‘presumed’
credibility (Önkal et al., 2017). Source credibility can be revised when forecast (in)accuracy is encountered
repetitively (Jiang et al., 1996), with forecaster and user confidence playing key roles (Sah et al., 2013).
Trust is critical for forecasting efforts to be translated into sound decisions (Choi et al., 2020; Özer
et al., 2011). Further work on fostering trust in individual/collaborative forecasting will benefit from how
trusted experts and models are selected and combined to enhance decision-making.

2.11. Evaluation, validation, and calibration

2.11.1. Point, interval, and pHDR forecast error measures75
Point forecasts are single number forecasts for an unknown future quantity also given by a single num-
ber. Interval forecasts take the form of two point forecasts, an upper and a lower limit. Finally, a less
common type of forecast would be a predictive Highest Density Region (pHDR), i.e., an HDR (Hyndman,
1996) for the conditional density of the future observable. pHDRs would be interesting for multimodal
(possibly implicit) predictive densities, e.g., in scenario planning. Once we have observed the correspond-
ing realisation, we can evaluate our point, interval and pHDR forecasts.
There are many common point forecast error measures (PFEMs), e.g., the mean squared error (MSE),
mean absolute error (MAE), mean absolute scaled error (MASE), mean absolute percentage error (MAPE)
or many others (see section 3.4 in Hyndman and Athanasopoulos, 2018). Which one is most appropriate
for our situation, or should we even use multiple different PFEMs?
Let us take a step back. Assume we have a full density forecast and wish to “condense” it to a point
forecast that will minimise some PFEM in expectation. The key observation is that different PFEMs will be
minimised by different point forecasts derived from the same density forecast (Kolassa, 2020b).

• The MSE is minimised by the expectation.

• The MAE and MASE are minimised by the median (Hanley et al., 2001).
• The MAPE is minimised by the (−1)-median (Gneiting, 2011a, p. 752 with β = −1).
• The hinge/tick/pinball loss is minimised by the appropriate quantile (Gneiting, 2011b).
• In general, there is no loss function that is minimised by the mode (Heinrich, 2014).

Our forecasting algorithm may not output an explicit density forecast. It is nevertheless imperative
to think about which functional of the implicit density we want to elicit (Gneiting, 2011a), and tailor our
error measure – and forecasting algorithm! – to it. It usually makes no sense to evaluate a point forecast
with multiple PFEMs (Kolassa, 2020b).

75 This subsection was written by Stephan Kolassa.

73
 Interval forecasts can be specified in multiple ways. We can start with a probability coverage and re-
quire two appropriate quantiles – e.g., we could require a 2.5% and a 97.5% quantile forecast, yielding a
symmetric or equal-tailed 95% interval forecast. Interval forecasts of this form can be evaluated by the
interval score (Winkler, 1972; Brehmer and Gneiting, 2020), a proper scoring rule (section 6.2 in Gneiting
and Raftery, 2007). We can also use the hinge loss to evaluate the quantile forecasts separately. Alter-
natively, we can require a shortest interval subject to a specified coverage. This interval is not elicitable
relative to practically relevant classes of distributions (Brehmer and Gneiting, 2020). Yet another possibil-
ity is to maximise the interval forecast’s probability coverage, subject to a maximum length 2`. This modal
interval is elicitable by an appropriate `-zero-one-loss (Brehmer and Gneiting, 2020).
The pHDR is not elicitable even for unimodal densities (Brehmer and Gneiting, 2020). In the multi-
modal case, the analysis is likely difficult. Nevertheless, a variation of the Winkler score has been proposed
to evaluate pHDRs on an ad hoc basis (Hyndman, 2020). One could also compare the achieved to the nom-
inal coverage, e.g., using a binomial test – which disregards the volume of the pHDR (Kolassa, 2020a).
In conclusion, here is a bewildering array of PFEMs, which require more thought in choosing among
than is obvious at first glance. The difficulties involved in evaluating interval and pHDR forecasts motivate
a stronger emphasis on full density forecasts (cf. Askanazi et al., 2018).

2.11.2. Evaluating probabilistic forecasts76

Probabilistic forecasting is a term that is not strictly defined, but usually refers to everything beyond
point forecasting (Gneiting, 2011a). However, in this section we consider only the evaluation of full predic-
tive distributions or equivalent characterisations. For the evaluation of prediction of quantiles, intervals
and related objects, see section 2.11.1.
One crucial point for evaluating probabilistic forecasts is the reporting, which is highly influenced from
meteorologic communities. From the theoretical point of view, we should always report the predicted cu-
mulative distribution function F
b of our prediction target F. Alternatively for continuous data, reporting
the probability density function is a popular choice. For univariate prediction problems a common alter-
native is to report quantile forecast on a dense grid of probabilities, as it approximates the full distribution
(Hong et al., 2016). For multivariate forecasts, it seems to become standard to report a large ensemble (a
set of simulated trajectories/paths) of the full predictive distribution. The reason is that the reporting of
a multivariate distribution (or an equivalent characterisation) of sophisticated prediction models is often
not feasible or practicable, especially for non-parametric or copula-based forecasting methods.
In general, suitable tools for forecasting evaluation are proper scoring rules as they address calibration
and sharpness simultaneously (Gneiting and Raftery, 2007; Gneiting and Katzfuss, 2014). Preferably, we
consider strictly proper scoring rules which can identify the true predicted distribution among a set of
forecast candidates that contains the true model.
In the univariate case the theory is pretty much settled and there is quite some consensus about the eval-
uation of probabilistic forecasts (Gneiting and Katzfuss, 2014). The continuous ranked probability score
(CRPS) and logarithmic scores (log-score) are popular strictly proper scoring rules, while the quadratic
and pseudospherical score remain strictly proper alternatives. The CRPS can be well approximated by
averaging across quantile forecasts on an equidistant grid of probabilities (Nowotarski and Weron, 2018).

76 This subsection was written by Florian Ziel.

74
 For multivariate forecast evaluation the situation is more complicated and many questions remain open
(Gneiting and Raftery, 2007; Meng et al., 2020). The multivariate version of the log-score is a strictly proper
scoring rule, but it requires the availability of a multivariate density forecast. This makes it impracticable
for many applications. Gneiting and Raftery (2007) discuss the energy score, a multivariate generalisation
of the CRPS, that is strictly proper. Still, it took the energy score more than a decade to increase its
popularity in forecasting. A potential reason is the limited simulation study of Pinson and Tastu (2013)
that concludes that the energy score can not discriminate well differences in the dependency structure.
In consequence other scoring rules were proposed in literature, e.g., the variogram score (Scheuerer and
Hamill, 2015) which is not strictly proper. Ziel and Berk (2019) consider a strictly proper scoring method
for continuous variables using copula techniques. In contrast to Pinson and Tastu (2013), recent studies
(Ziel and Berk, 2019; Lerch et al., 2020) show that the energy score discriminates well when used together
with significance tests like the Diebold-Mariano (DM) test.
In general, we recommended scoring be applied with reliability evaluation (see section 2.11.3) and
significance tests (see section 2.11.4). Additionally, if we want to learn about the performance of our
forecasts it is highly recommended to consider multiple scoring rules and evaluate on lower-dimensional
subspaces. For multivariate problems, this holds particularly for the evaluation of univariate and bivariate
marginal distributions.

2.11.3. Assessing the reliability of probabilistic forecasts77

A probabilistic forecast takes the form of a predictive distribution over future quantities or events of
interest. Probabilistic forecasts are central in risk-based decision making where reliability, or calibration,
is a necessary condition for the optimal use and value of the forecast. A probabilistic forecast is calibrated
if the observation cannot be distinguished from a random draw from the predictive distribution or, in the
case of ensemble forecasts, if the observation and the ensemble members look like random draws from the
same distribution.
In the univariate setting, several alternative notions of calibration exist for both a single forecast (Gneit-
ing et al., 2007; Tsyplakov, 2013) and a group of forecasts (Strähl and Ziegel, 2017). The notion most
commonly used in applications is probabilistic calibration (Dawid, 1984); the forecast system is proba-
bilistically calibrated if the probability integral transform (PIT) of a random observation, that is, the value
of the predictive cumulative distribution function in the observation, is uniformly distributed. If the pre-
dictive distribution has a discrete component, a randomised version of the PIT should be used (Gneiting
and Ranjan, 2013).
Probabilistic calibration is assessed visually by plotting the histogram of the PIT values over a test set.
A calibrated forecast system will return a uniform histogram, a ∩-shape indicates overdispersion and a ∪-
shape indicates underdispersion, while a systematic bias results in a biased histogram (e.g. Thorarinsdottir
and Schuhen, 2018). The discrete equivalent of the PIT histogram, which applies to ensemble forecasts, is
the verification rank histogram (Anderson, 1996; Hamill and Colucci, 1997). It shows the distribution of
the ranks of the observations within the corresponding ensembles and has the same interpretation as the
PIT histogram.

77 This subsection was written by Thordis Thorarinsdottir.

75
 For small test sets, the bin number of a PIT/rank histogram must be chosen with care. With very
few bins, the plot may obscure miscalibration while with many bins, even perfectly calibrated forecasts
can yield non-uniformly appearing histograms (Thorarinsdottir and Schuhen, 2018; Heinrich, 2020). The
bin number should be chosen based on the size of the test set, with the bin number increasing linearly
with the size of the test set (Heinrich, 2020). More specifically, the uniformity of PIT/rank values can be
assessed with statistical tests (Delle Monache et al., 2006; Taillardat et al., 2016; Wilks, 2019), where the
test statistics can be interpreted as a distance between the observed and a flat histogram (Wilks, 2019;
Heinrich, 2020). Testing predictive performance is further discussed in section 2.11.4.
Calibration assessment of multivariate forecasts is complicated by the lack of a unique ordering in
higher dimensions and the many ways in which the forecasts can be miscalibrated (Wilks, 2019). Gneiting
et al. (2008) propose a general two-step approach where an ensemble forecast and the corresponding ob-
servation are first mapped to a single value by a pre-rank function. Subsequently, the pre-rank function
values are ranked in a standard manner. The challenge here is to find a pre-rank function that yields in-
formative and discriminative ranking (Wilks, 2004; Gneiting et al., 2008; Thorarinsdottir et al., 2016), see
Thorarinsdottir et al. (2016) and Wilks (2019) for comparative studies. Alternatively, Ziegel and Gneiting
(2014) propose a direct multivariate extension of the univariate setting based on copulas.
In addition to calibration, forecasts should also be sharp, or specific, to ensure their utility in decision
making (e.g. Gneiting et al., 2007). This is discussed in sections 2.11.1 and 2.11.2.

2.11.4. Statistical tests of forecast performance78

The rich domain of forecasting has created numerous methods for generating competing forecasts. A
natural consequence of such was the development of statistical tests for predictive ability in the last thirty
years. These tests provided forecasters some formal reassurance that the predictive superiority of a leading
forecast is statistically significant and is not merely due to random chance.
One of the early papers that undoubtedly sparked growth in this field was Diebold and Mariano (1995,
DM hereafter). In their seminal paper, DM provided a simple yet general approach for testing equal predic-
tive ability, i.e., if two forecasting sources are equally accurate on average. Their population-level predic-
tive ability test has very few assumptions (e.g., covariance stationary loss differential) and is applicable to
a wide range of loss functions, multi-period settings, and wide class of forecast errors (e.g., non-Gaussian,
serially and/or contemporaneously correlated). This test though not originally intended for models has
been widely used by others to test forecasting models’ accuracy (Diebold, 2015).
Modifications were later introduced by Harvey et al. (1998) to improve small sample properties of the
test. Generalisations and extensions have emerged to address issues that DM tests encountered in practice
such as nested models (Clark and McCracken, 2001, 2009), parameter estimation error (West, 1996), coin-
tegrated variables (Corradi et al., 2001), high persistence (Rossi, 2005), and panel data (Timmermann and
Zhu, 2019). Finite-sample predictive ability tests also emerged from the observation that models may have
equal predictive ability in finite samples, which generated a class called conditional predictive accuracy
tests (Giacomini and White, 2006; Clark and McCracken, 2013).
An alternative approach to comparing forecast accuracy is through the notion of forecast encompassing,
which examines if a forecast encompasses all useful information from another with respect to predictions

78 This subsection was written by Victor Richmond R. Jose.

76
(Chong and Hendry, 1986; Harvey et al., 1998; Clark and McCracken, 2001). Though it has a few more
assumptions, forecast encompassing tests in certain contexts might be preferable to the mean square pre-
diction error tests à la Diebold-Mariano (Busetti and Marcucci, 2013).
Another stream of statistical tests that are available looks at multiple forecasts simultaneously instead
of pairs. Addressing a need for a reality check on “data snooping”, White (2000) later modified by Hansen
(2005) developed a multiple model test that uses a null hypothesis of “superior predictive ability” instead
of the equal predictive ability used in DM tests. These have also been generalised to deal with issues such
as cointegrated variables (Corradi et al., 2001) and multi-horizon forecasts (Quaedvlieg, 2019). Recently,
Li et al. (2020e) proposed a conditional superior predictive ability test similar to Giacomini and White
(2006)’s innovation to the DM test. A different approach for studying performance of multiple forecast-
ing models is through the use of multiple comparison tests such as multiple comparison with a control
and multiple comparison with the best (Hsu, 1981; Edwards and Hsu, 1983; Horrace and Schmidt, 2000).
These tests often are based on jointly estimated confidence intervals that measure the difference between
two parameters of interest such as the forecast accuracies of a model and a benchmark. Koning et al.
(2005) illustrates how they can be ex post used to analyse forecasting performance in the M3 forecasting
competition (Makridakis and Hibon, 2000) using model ranking instead of forecast accuracy scores as its
primitives. The multiple comparison of the best was used in the analysis of the subsequent M4 and M5
Competitions (Makridakis et al., 2020c,d).

2.12. The future of forecasting theory79

The theory of forecasting appears mature today, based on dedicated developments at the interface
among a number of disciplines, e.g., mathematics and statistics, computer sciences, psychology, etc. A
wealth of these theoretical developments have originated from specific needs and challenges in different
application areas, e.g., in economics, meteorology and climate sciences, as well as management science
among others. In this section, many aspects of the theory of forecasting were covered, with aspects related
to data, modelling and reasoning, forecast verification. Now, the fact that forecasting is mature does not
mean that all has been done – we aim here at giving a few pointers at current and future challenges.
First of all, it is of utmost importance to remember that forecasting is a process that involves both
quantitative aspects (based on data and models) and humans, at various levels, i.e., from the generation
of forecasts to their use in decision-making. A first consequence is that we always need to find, depend-
ing on the problem at hand, an optimal trade-off between data-driven approaches and the use of expert
judgement. In parallel, forecasting is to be thought of in a probabilistic framework in a systematic manner
(Gneiting and Katzfuss, 2014). This allows us to naturally convey uncertainty about the future, while pro-
viding the right basis to make optimal decisions in view of the characteristics of the decision problem, as
well as the loss (or utility) function and risk aversion of the decision maker. Another consequence is that
using forecasts as input to decision-making often affects the outcome to be predicted itself – a problem
known as self-negating forecasts (possibly also self-fulfilling) or the prophet dilemma. With advances in
the science of dynamic systems and game theory, we should invest in modelling those systems as a whole
(i.e., forecasting and decision-making) in order to predict the full range of possible outcomes, based on the
decisions that could be made.

79 This subsection was written by Pierre Pinson.

77
 In parallel, it is clear that today, the amount of data being collected and possibly available for fore-
casting is growing at an astounding pace. This requires re-thinking our approaches to forecasting towards
high-dimensional models, online learning, etc. Importantly, the data being collected is distributed in terms
of ownership. And, due to privacy concerns and competitive interests, some may not be ready to share their
data. Novel frameworks to learning and forecasting ought to be developed with that context in mind, for
instance focusing on distributed and privacy-preserving learning – an example among many others is that
of Google pushing forward federated learning (Abadi et al., 2016), an approach to deep learning where the
learning process is distributed and with a privacy layer. Eventually the access and use of data, as well as
the contribution to distributed learning (and collaborative analytics, more generally), may be monetised,
bringing a mechanism design component to the future theory of forecasting. A simple and pragmatic ex-
ample is that of forecast reconciliation: if asking various agents to modify their forecasts to make them
coherent within a hierarchy, such modifications could be monetised to compensate for accuracy loss.
A large part of today’s modelling and forecasting approaches uses a wealth of data to identify and
fit models, to be eventually used to forecast based on new data and under new conditions. Different
approaches have been proposed to maximise the generalisation ability of those models, to somewhat max-
imise chances to do well out-of-sample. At the root of this problem is the effort to go beyond correlation
only, and to identify causality (see, e.g., Pearl (2009) for a recent extensive coverage). While causality has
been a key topic of interest to forecasters for a long time already, new approaches and concepts are being
pushed forward for identification of and inference in causal models (Peters et al., 2017), which may have a
significant impact on the theory of forecasting.
Eventually, the key question of what a good forecast is will continue to steer new developments in the
theory of forecasting in the foreseeable future. The nature of goodness of forecasts (seen from the meteo-
rological application angle) was theorised a few decades ago already (Murphy, 1993), based on consistency,
quality and value. We still see the need to work further on that question – possibly considering these 3 pil-
lars, but possibly also finding other ways to define desirable properties of forecasts. This will, in all cases,
translates to further developing frameworks for forecast verification, focusing on the interplay between
forecast quality and value, but also better linking to psychology and behavioural economics. In terms of
forecast verification, some of the most pressing areas most likely relate to (multivariate) probabilistic fore-
casting and to the forecasting of extreme events. When it comes to forecast quality and value, we need to
go beyond the simple plugging of forecasts into decision problems to assess whether this yields better de-
cisions, or not. Instead, we ought to propose suitable theoretical frameworks that allow assessing whether
certain forecasts are fundamentally better (than others) for given classes of decision problems. Finally, the
link to psychology and behavioural economics should ensure a better appraisal of how forecasts are to be
communicated, how they are perceived and acted upon.
Most of the advances in the science of forecasting have come from the complementarity between theo-
retical developments and applications. We can then only be optimistic for the future since more and more
application areas are relying heavily on forecasting. Their specific needs and challenges will continue
fuelling upcoming developments in the theory of forecasting.

78
3. Practice

3.1. Introduction to forecasting practice80

The purpose of forecasting is to improve decision making in the face of uncertainty. To achieve this,
forecasts should provide an unbiased guess at what is most likely to happen (the point forecast), along with
a measure of uncertainty, such as a prediction interval (PI). Such information will facilitate appropriate
decisions and actions.
Forecasting should be an objective, dispassionate exercise, one that is built upon facts, sound reasoning,
and sound methods. But since forecasts are created in social settings, they are influenced by organisational
politics and personal agendas. As a consequence, forecasts will often reflect aspirations rather than unbi-
ased projections.
In organisations, forecasts are created through processes that can involve multiple steps and partici-
pants. The process can be as simple as executive fiat (also known as evangelical forecasting), unencum-
bered by what the data show. More commonly, the process begins with a statistical forecast (generated by
forecasting software), which is then subject to review and adjustment, as illustrated in figure 4.

Figure 4: Multi-stage forecasting process.

In concept, such an elaborate multi-stage process allows “management intelligence” to improve forecast
quality, incorporating information not accounted for in the statistical model. In reality, however, benefits
are not assured. Lawrence et al. (2006) reviewed more than 200 studies, concluding that human judgment
can be of significant benefit but is also subject to significant biases. Among the many papers on this subject,
there is general agreement on the need to track and review overrides, and the need to better understand
the psychological issues around judgmental adjustments.
The underlying problem is that each human touch point subjects the forecast to the interests of the re-
viewers – and these interests may not align with creating an accurate, unbiased forecast. To identify where
such problems are occurring, Forecast Value Added (FVA) analysis is an increasingly popular approach
among practitioners.
FVA is defined as the change in a forecasting performance metric that can be attributed to a particular
step or participant in the forecasting process (Gilliland, 2002). Any activity that fails to deliver positive
FVA (i.e., fails to improve forecast quality) is considered process waste.
Starting with a naive forecast, FVA analysis seeks to determine whether each subsequent step in the
process improves upon the prior steps. The “stairstep report” of table 1 is a familiar way of summarising
results, as in this example from Newell Rubbermaid (Schubert and Rickard, 2011).
Here, averaged across all products, naive (random walk) achieved forecast accuracy of 60%. The com-
pany’s statistical forecast delivered five percentage points of improvement, but management review and
adjustment delivered negative value. Such findings – not uncommon – urge further investigation into

80 This subsection was written by Michael Gilliland.

79
 Table 1: Stairstep report showing FVA results.

Process Step Forecast accuracy FVA vs. Naive FVA vs. Statistical
(100%−MAPE)
Naive forecast 60%
Statistical forecast 65% 5%
Adjusted forecast 62% 2% -3%

causes and possible process corrections (such as training reviewers or limiting adjustments). Alternatively,
the management review step could be eliminated, providing the dual benefits of freeing up management
time spent on forecasting and, on average, more accurate forecasts.
Morlidge (2014c) expanded upon FVA analysis to present a strategy for prioritising judgmental ad-
justments, finding the greatest opportunity for error reduction in products with high volume and high
relative absolute error. Chase (2021) described a machine learning (ML) method to guide forecast review,
identifying which forecasts are most likely to benefit from adjustment along with a suggested adjustment
range. Baker (2021) used ML classification models to identify characteristics of non-value adding over-
rides, proposing the behavioural economics notion of a “nudge” to prompt desired forecaster behaviour.
Further, Goodwin et al. (2017) derived upper bounds for FVA relative to naive forecasts. And de Kok (2017)
created a Stochastic Value Added (SVA) metric to assess the difference between actual and forecasted dis-
tributions, knowledge of which is valuable for inventory management.
Including an indication of uncertainty around the point forecast remains an uncommon practice. Pre-
diction intervals in software generally underestimate uncertainty, often dramatically, leading to unrealistic
confidence in the forecast. And even when provided, PIs largely go unused by practitioners. Goodwin
(2014) summarised the psychological issues, noting that the generally poor calibration of the PIs may not
explain the reluctance to utilise them. Rather, “an interval forecast may accurately reflect the uncertainty,
but it is likely to be spurned by decision makers if it is too wide and judged to be uninformative” (Goodwin,
2014, page 5).
It has long been recognised (Chatfield, 1986; Lawrence, 2000) that the practice of forecasting falls well
short of the potential exhibited in academic research, and revealed by the M forecasting competitions. In
the M4, a simple benchmark combination method (the average of Single, Holt, and Damped exponential
smoothing) reduced the overall weighted average (OWA) error by 17.9% compared to naive. The top six
performing methods in M4 further reduced OWA by over 5% compared to the combination benchmark
(Makridakis et al., 2020c). But in forecasting practice, just bettering the accuracy of naive has proven to be
a surprising challenge. Morlidge’s (2014b) study of eight consumer and industrial businesses found 52%
of their forecasts failed to do so. And, as shown, Newel Rubbermaid beat naive by just two percentage
points after management adjustments.
Ultimately, forecast accuracy is limited by the nature of the behaviour being forecast. But even a highly
accurate forecast is of little consequence if overridden by management and not used to enhance decision
making and improve organisational performance.
Practitioners need to recognise limits to forecastability and be willing to consider alternative (non-
forecasting) approaches when the desired level of accuracy is not achievable (Gilliland, 2010). Alternatives
include supply chain re-engineering – to better react to unforeseen variations in demand, and demand

80
smoothing – leveraging pricing and promotional practices to shape more favourable demand patterns.
Despite measurable advances in our statistical forecasting capabilities (Makridakis et al., 2020b), it is
questionable whether forecasting practice has similarly progressed. The solution, perhaps, is what Mor-
lidge (2014a, page 39) suggests that “users should focus less on trying to optimise their forecasting process
than on detecting where their process is severely suboptimal and taking measures to redress the problem”.
This is where FVA can help.
For now, the challenge for researchers remains: To prompt practitioners to adopt sound methods based
on the objective assessment of available information, and avoid the “worst practices” that squander re-
sources and fail to improve the forecast.

3.2. Operations and supply chain management

3.2.1. Demand management81
Demand management is one of the dominant components of supply chain management (Fildes et al.,
2006). Accurate demand estimate of the present and future is a first vital step for almost all aspects of
supply chain optimisation, such as inventory management, vehicle scheduling, workforce planning, and
distribution and marketing strategies (Kolassa and Siemsen, 2016). Simply speaking, better demand fore-
casts can yield significantly better supply chain management, including improved inventory management
and increased service levels. Classic demand forecasts mainly rely on qualitative techniques, based on ex-
pert judgement and past experience (e.g., Weaver, 1971), and quantitative techniques, based on statistical
and machine learning modelling (e.g., Taylor, 2003b; Bacha and Meyer, 1992). A combination of qualita-
tive and quantitative methods is also popular and proven to be beneficial in practice by, e.g., judgemental
adjustments (Turner, 1990; Önkal and Gönül, 2005; Syntetos et al., 2016b), judgemental forecast model
selection (Petropoulos et al., 2018b; Han et al., 2019), and other advanced forecasting support systems
(Baecke et al., 2017; Arvan et al., 2019).
The key challenges that demand forecasting faces vary from domain to domain. They include:

1. The existence of intermittent demands, e.g., irregular demand patterns of fashion products. Ac-
cording to Nikolopoulos (2020), limited literature has focused on intermittent demand. The seminal
work by Croston (1972) was followed by other representative methods such as the SBA method by
Syntetos and Boylan (2001), the aggregate–disaggregate intermittent demand approach (ADIDA) by
Nikolopoulos et al. (2011), the multiple temporal aggregation by Petropoulos and Kourentzes (2015),
and the k nearest neighbour (kNN) based approach by Nikolopoulos et al. (2016). See section 2.7 for
more details on intermittent demand forecasting.
2. The emergence of new products. Recent studies on new product demand forecasting are based on
finding analogies (Wright and Stern, 2015; Hu et al., 2019), leveraging comparable products (Baard-
man et al., 2018), and using external information like web search trends (Kulkarni et al., 2012). See
section 3.2.7 for more details on new product demand forecasting.
3. The existence of short-life-cycle products, e.g., smartphone demand (e.g., Szozda, 2010; Chung et al.,
2012; Shi et al., 2020).

81 This subsection was written by Yanfei Kang.

81
 4. The hierarchical structure of the data such as the electricity demand mapped to a geographical hier-
archy (e.g., Athanasopoulos et al., 2009; Hyndman et al., 2011; Hong et al., 2019, see section 2.9.1).

With the advent of the big data era, a couple of coexisting new challenges have drawn the attention of
researchers and practitioners in the forecasting community: the need to forecast a large volume of related
time series (e.g., thousands or millions of products from one large retailer: Salinas et al., 2019a), and the
increasing number of external variables that have significant influence on future demand (e.g., massive
amounts of keyword search indices that could impact future tourism demand (Law et al., 2019)). Recently,
to deal with these new challenges, numerous empirical studies have identified the potentials of deep learn-
ing based global models, in both point and probabilistic demand forecasting (e.g., Wen et al., 2017; Ran-
gapuram et al., 2018; Salinas et al., 2019a; Bandara et al., 2020b). With the merits of cross-learning, global
models have been shown to be able to learn long memory patterns and related effects (Montero-Manso and
Hyndman, 2020), latent correlation across multiple series (Smyl, 2020), handle complex real-world fore-
casting situations such as data sparsity and cold-starts (Chen et al., 2020), include exogenous covariates
such as promotional information and keyword search indices (Law et al., 2019), and allow for different
choices of distributional assumptions (Salinas et al., 2019a).

3.2.2. Forecasting in the supply chain82

A supply chain is ‘a network of stakeholders (e.g., retailers, manufacturers, suppliers) who collaborate
to satisfy customer demand’ (Perera et al., 2019). Forecasts inform many supply chain decisions, includ-
ing those relating to inventory control, production planning, cash flow management, logistics and human
resources. Typically, forecasts are based on an amalgam of statistical methods and management judgment
(Fildes and Goodwin, 2007). Hofmann and Rutschmann (2018) have investigated the potential for using
big data analytics in supply chain forecasting but indicate more research is needed to establish its useful-
ness.
In many organisations forecasts are a crucial element of Sales and Operations Planning (S&OP), a tool
that brings together different business plans, such as those relating to sales, marketing, manufacturing
and finance, into one integrated set of plans (Thomé et al., 2012). The purposes of S&OP are to balance
supply and demand and to link an organisation’s operational and strategic plans. This requires collab-
oration between individuals and functional areas at different levels because it involves data sharing and
achieving a consensus on forecasts and common objectives (Mello, 2010). Successful implementations of
S&OP are therefore associated with forecasts that are both aligned with an organisation’s needs and able
to draw on information from across the organisation. This can be contrasted with the ‘silo culture’ iden-
tified in a survey of companies by Moon et al. (2003) where separate forecasts were prepared by different
departments in ‘islands of analysis’. Methods for reconciling forecasts at different levels in both cross-
sectional hierarchies (e.g., national, regional and local forecasts) and temporal hierarchies (e.g., annual,
monthly and daily forecasts) are also emerging as an approach to break through information silos in or-
ganisations. Cross-temporal reconciliation provides a data-driven approach that allows information to
be drawn from different sources and levels of the hierarchy and enables this to be blended into coherent
forecasts (Kourentzes and Athanasopoulos, 2019).

82 This subsection was written by Paul Goodwin.

82
 In some supply chains, companies have agreed to share data and jointly manage planning processes
in an initiative known as Collaborative Planning, Forecasting, and Replenishment (CPFR) (Seifert, 2003).
CPFR involves pooling information on inventory levels and on forthcoming events, like sales promotions.
Demand forecasts can be shared, in real time via the Internet, and discrepancies between them reconciled.
In theory, information sharing should reduce forecast errors. This should mitigate the ‘bullwhip effect’
where forecast errors at the retail-end of supply chains cause upstream suppliers to experience increasingly
volatile demand, forcing them to hold high safety stock levels (Lee et al., 2007).
Much research demonstrating the benefits of collaboration has involved simulated supply chains (Fildes,
2017). Studies of real companies have also found improved performance through collaboration (e.g., Boone
and Ganeshan, 2008; Hill et al., 2018; Eksoz et al., 2019), but case study evidence is still scarce (Syntetos
et al., 2016a). The implementation of collaborative schemes has been slow with many not progressing be-
yond the pilot stage (Panahifar et al., 2015; Galbreth et al., 2015). Barriers to successful implementation
include a lack of trust between organisations, reward systems that foster a silo mentality, fragmented fore-
casting systems within companies, incompatible systems, a lack of relevant training and the absence of top
management support (Fliedner, 2003; Thomé et al., 2014).
Initiatives to improve supply chain forecasting can be undermined by political manipulation of fore-
casts and gaming. Examples include ‘enforcing’: requiring inflated forecasts to align them with sales or
financial goals, ‘sandbagging’: underestimating sales so staff are rewarded for exceeding forecasts, and
‘spinning’: manipulating forecasts to garner favourable reactions from colleagues (Mello, 2009). Pennings
et al. (2019) discuss schemes for correcting such intentional biases.

3.2.3. Forecasting for inventories83

Three aspects of the interaction between forecasting and inventory management have been studied in
some depth and are the subject of this review: the bullwhip effect, forecast aggregation, and performance
measurement.
The ‘bullwhip effect’ occurs whenever there is amplification of demand variability through the supply
chain (Lee et al., 2004), leading to excess inventories. This can be addressed by supply chain members
sharing downstream demand information, at stock keeping unit level, to take advantage of less noisy data.
Analytical results on the translation of ARIMA demand processes have been established for order-up-to
inventory systems (Gilbert, 2005). There would be no value in information sharing if the wholesaler can
use such relationships to deduce the retailer’s demand process from their orders (see, for example, Graves,
1999). Such deductions assume that the retailer’s demand process and demand parameters are common
knowledge to supply chain members. Ali and Boylan (2011) showed that, if such common knowledge
is lacking, there is value in sharing the demand data itself and Ali et al. (2012) established relationships
between accuracy gains and inventory savings. Analytical research has tended to assume that demand
parameters are known. Pastore et al. (2020) investigated the impact of demand parameter uncertainty,
showing how it exacerbates the bullwhip effect.
Forecasting approaches have been developed that are particularly suitable in an inventory context, even
if not originally proposed to support inventory decisions. For example, Nikolopoulos et al. (2011) proposed
that forecasts could be improved by aggregating higher frequency data into lower frequency data. Forecasts

83 This subsection was written by John E. Boylan.

83
are generated at the lower frequency level and then disaggregated, if required, to the higher frequency
level. For inventory replenishment decisions, the level of aggregation may conveniently be chosen to be
the lead time, thereby taking advantage of the greater stability of data at the lower frequency level, with
no need for disaggregation.
The variance of forecast errors over lead time is required to determine safety stock requirements for
continuous review systems. The conventional approach is to take the variance of one-step-ahead errors
and multiply it by the lead time. However, this estimator is unsound, even if demand is independent and
identically distributed, as explained by Prak et al. (2017). A more direct approach is to smooth the mean
square errors over the lead time (Syntetos and Boylan, 2006).
Strijbosch and Moors (2005) showed that unbiased forecasts will not necessarily lead to achievement,
on average, of target cycle service levels or fill rates. Wallström and Segerstedt (2010) proposed a ‘Periods
in Stock’ measure, which may be interpreted, based on a ‘fictitious stock’, as the number of periods a unit
of the forecasted item has been in stock or out of stock. Such measures may be complemented by a detailed
examination of error-implication metrics (Boylan and Syntetos, 2006). For inventory management, these
metrics will typically include inventory holdings and service level implications (e.g., cycle service level, fill
rate). Comparisons may be based on total costs or via ‘exchange curves’, showing the trade-offs between
service and inventory holding costs. Comparisons such as these are now regarded as standard in the
literature on forecasting for inventories and align well with practice in industry.

3.2.4. Retail sales forecasting84

Retail companies depend crucially on accurate demand forecasting to manage their supply chain and
make decisions concerning planning, marketing, purchasing, distribution, and labour force (Ma and Fildes,
2017). Inaccurate forecasts lead to unnecessary costs and poor customer satisfaction. Inventory levels
should be neither too high to avoid waste and extra costs (of storage and labour force) that reduce sales
profits, nor too low to prevent stock-outs and sales lost. Forecasting product demand in retail can be seen
in a three-dimensional space: the position in the supply chain hierarchy (store, distribution centre, or
chain), the level in the product hierarchy (SKU, brand, category, or total) and the time granularity (day,
week, month, quarter, or year); see Syntetos et al. (2016a). In general, the higher is the position in the
supply chain, the lower is the time granularity.
Typically, retailers need daily forecasts for store replenishment and weekly forecasts for the distribution
centre (DC) activities at the SKU-level (Fildes et al., 2019b). Hierarchical forecasting is a quite promising
tool to generate coherent demand forecasts on multiple levels over different dimensions (Oliveira and
Ramos, 2019). Several factors may affect retail sales at the product-level. Data often exhibit multiple
seasonal patterns namely weekly and annual cycles which can be modelled by Fourier terms or dummies
(Huang et al., 2019). Demand increases substantially during holidays, festivals, and other special events
such as Christmas and Easter which can be captured by dummies. Weather conditions may also affect sales
namely high temperature. In contrast to calendar events whose dates are known weather variables need to
be forecasted. Price reductions and feature advertisement can boost sales considerably. Different types of
promotions are used by retailers such as free units for multiple purchases, direct discounts, loyalty card
bonuses, etc. A promotion on one product may affect both the sales of competing products and its own

84 This subsection was written by Patrı́cia Ramos.

84
futures sales. The high dimensionality of the explanatory variable space makes product sales difficult to
forecast using traditional approaches since overfitting may easily occur or models cannot even be estimated
(Trapero et al., 2015).
Univariate forecasting models are the most basic methods retailers may use to forecast demand at the
SKU level. They range from simple techniques such as simple moving averages or exponential smoothing
to Fourier analysis or ARIMA models. These methods should only be used to forecast demand at higher
aggregation levels or to forecast products with no promotional intensity (Ramos et al., 2015; Ramos and
Oliveira, 2016).
Regression-type models allow the inclusion of external effects that may explain some of the variation
of products sales (van Donselaar et al., 2016; Gur Ali and Pinar, 2016). Huang et al. (2014) and Ma et al.
(2016) developed successfully autoregressive distributed lag (ADL) models to forecast SKU retail sales
from many different categories conducting variable selection and estimation using LASSO. Arunraj and
Ahrens (2015) proposed a SARIMAX model to forecast daily sales of fresh produce. There is no clear
evidence yet that nonlinear models and novel machine learning methods can improve forecast accuracy
(Fildes et al., 2019b).
Zero sales due to stock-outs or low demand are very often at the SKU × store level either at weekly
or daily granularity. Methods suitable for forecasting intermittent data (see section 2.7) have been widely
used in spare part sales forecasting but have not been not yet evaluated in the retail context. Also, the
effective use of forecasts should couple point estimates with quantile predictions or prediction intervals
that are particularly important for determining safety stock amounts needed for replenishment. However,
this is an under-investigated aspect of retail forecasting (Taylor, 2007; Kolassa, 2016).

3.2.5. Online retail forecasting85

Online sales account for an ever-increasing proportion of retail sales, posing unique challenges to fore-
casting. There is not much specifically relevant literature, so we will call for more research and collect
aspects that should be kept in mind when forecasting online retail demand.
Many aspects are similar between brick and mortar (B&M) and online retail forecasting (cf. Fildes et al.,
2019b): demand is driven by multiple seasonalities and product lifecycles. Promotions are important in
both cases.
However, differences are important. First, online retailing is not monolithic. There are huge retailers
with enormous assortments and “Long Tails” of highly intermittent SKUs (even though they have fewer
large distribution centres than B&Ms have stores). Conversely, there are many small online retailers with
much smaller assortments. Here, turnovers may be so small that a large investment in forecasting is not
worthwhile.
Forecast use strongly depends on the retailer’s omnichannel strategy (Armstrong, 2017; Sopadjieva
et al., 2017; Melacini et al., 2018): e.g., for “order online, collect in store”, we need forecasts for both online
demand and for the demand fulfilled at the store. This and similar use cases can be addressed through
hierarchical forecasting with a large location hierarchy.
Online retailers, especially in fashion, have a much bigger problem with product returns. They may
need to forecast how many products are returned overall (e.g., Shang et al., 2020), or whether a specific

85 This subsection was written by Stephan Kolassa.

85
customer will return a specific product.
Online retailers can much more easily fine-tune interactions with the customer, e.g., the landing page,
product recommendations, personalised promotions, pricing or other offers, or other communications.
For this, retailers leverage the customer’s purchasing, browsing or returns history, current shopping cart
contents, or the retailer’s stock position in particular products. For instance, online retailers frequently
inform customers if stocks run low, which may either reduce (Park et al., 2020) or increase demand (Cui
et al., 2019). On the other hand, product reviews are a type of interaction of the customer with the retailer
which drives future demand and should be included in forecasts.
Most importantly, B&M retailers decouple pricing/promotion decisions and optimisation from the cus-
tomer interaction, and therefore from forecasting. Online, this is not possible, because the customer has
total transparency to competitors’ offerings. Thus, online pricing needs to react much more quickly to
competitive pressures – faster than the forecasting cycle. Thus, the specific value of demand drivers is
usually not known at the time of forecasting: we don’t know yet which customer will log on to our online
store, so we do not know yet how many people will see a particular product displayed on their personalised
landing page. (Nor do we know today what remaining stock will be displayed to the customer, per above.)
Thus, changes in drivers needs to be “baked into” the forecasting algorithm – but not even the coupling
between the promotional or interaction strategy and forecasting is obvious.
Feedback loops between forecasting and other processes are thus even more important online: yes-
terday’s forecasts drive today’s stock position, driving today’s personalised recommendations, driving de-
mand, driving today’s forecasts for tomorrow. Overall, online retail forecasting needs to be more agile and
responsive to the latest interactional decisions taken in the web store, and more tightly integrated into the
retailer’s interactional tactics and omnichannel strategy.
We are unaware of systematic research on forecasting online retail demand that addresses the points
above.

3.2.6. Promotional forecasting86

Promotional forecasting is central for retailing, but also relevant for many manufacturers, particularly
of Fast Moving Consumer Goods (FMCG). In principle, the objective is to forecast sales, as in most business
forecasting cases. However, what sets promotional forecasting apart is that we also make use of information
about promotional plans, pricing, and sales of complementary and substitute products (Bandyopadhyay,
2009; Zhang et al., 2008). Other relevant variables may include store location and format, variables that
capture the presentation and location of a product in a store, proxies that characterise the competition, and
so on (Van Heerde et al., 2002; Andrews et al., 2008).
Three modelling considerations guide us in the choice of models. First, promotional (and associated
effects) are proportional. For instance, we do not want to model the increase in sales as an absolute number
of units, but instead, as a percentage uplift. We do this to not only make the model applicable to both
smaller and larger applications, for example, small and large stores in a retailing chain, but also to gain a
clearer insight into the behaviour of our customers. Second, it is common that there are synergy effects.
For example, a promotion for a product may be offset by promotions for substitute products. Both these
considerations are easily resolved if we use multiplicative regression models. However, instead of working

86 This subsection was written by Nikolaos Kourentzes.

86
with the multiplicative models, we rely on the logarithmic transformation of the data and proceed to
construct the promotional model using the less cumbersome additive formulation. Third, the objective of
promotional models does not end with providing accurate predictions. We are also interested in the effect
of the various predictors: their elasticity. This can in turn provide the users with valuable information
about the customers, but also be an input for constructing optimal promotional and pricing strategies
(Zhang et al., 2008).
Promotional models have been widely used on brand-level data (for example, Divakar et al., 2005).
However, they are increasingly used on Stock Keeping Unit (SKU) level data (Trapero et al., 2015; Ma
et al., 2016), given advances in modelling techniques. Especially at that level, limited sales history and
potentially non-existing examples of past promotions can be a challenge. (Trapero et al., 2015) consider
this problem and propose using a promotional model that has two parts that are jointly estimated. The
first part focuses on the time series dynamics and is modelled locally for each SKU. The second part tack-
les the promotional part, which pools examples of promotions across SKUs to enable providing reasonable
estimates of uplifts even for new SKUs. To ensure the expected heterogeneity in the promotional effects,
the model is provided with product group information. Another recent innovation is looking at modelling
promotional effects both at the aggregate brand or total sales level, and disaggregate SKU level, relying
on temporal aggregation (Kourentzes and Petropoulos, 2016). Ma et al. (2016) concern themselves with
the intra-and inter-category promotional information. The challenge now is the number of variables to
be considered for the promotional model, which they address by using sequential LASSO. Although the
aforementioned models have shown very promising results, one has to recognise that in practice promo-
tions are often forecasted using judgemental adjustments, with inconsistent performance (Trapero et al.,
2013).

3.2.7. New product forecasting87

Obtaining an accurate sales forecast for new products is crucial to the successful introduction of these
products to the market, which in turn decides the survivability of the organisations (Goodwin et al., 2014).
Traditionally forecasting is about finding an underlying pattern in the past data and projecting that into the
future. Obviously, for a new product, this mechanism would not work as there is no past data.88 Assuming
the new product is totally new, to forecast in such situations forecasters historically resorted to qualitative
methods – which come with inevitable biases and the forecasts produced invariably are off the mark with
the forecast error that is seldom anywhere near acceptable.
Common-sense logic suggests that any human decision, including sales forecasting of a new product,
relies on two strands of information, namely what has happened in the past and what is expected to happen
in the future. Although it is acknowledged that there is no past data for the new product, there exists a vast
amount of other past data which includes observations of sales patterns for various products, which lead
to the realisation that sales of most products follows a general growth pattern. Such patterns are similar
in their features for most products, although the size of the features could differ from product to product.
Such patterns showing the sales growth are formally known as ‘growth curves/diffusion models’, the pearl
(simple logistic) curve and Gompertz curve being examples. These models are further enriched by adding

87 This subsection was written by Sheik Meeran.

88 The newness could be in quite varying levels, needing different treatments with varying difficulties.

87
factors that could explain the growth and in turn, the shape of the growth curves. The parameters in the
growth curves operationalise such mechanism. One such work is promulgated by Bass (1969), which is
extended multiple times to refine and enrich the growth-curve model in order to reflect the actual growth
closely (Islam and Meade, 2012). For more details on the Bass model (BM) and growth curves, please refer
to sections 2.3.19 and 2.3.20.
Through these growth models, the complexity of the task of new product forecasting is reduced to just
identifying a historic product whose growth pattern is likely to be followed by the new product. Such
product in forecasting parlance is called as ‘an analogous product’. In modelling terms, the product will
have a growth curve defined by the same parameters as that of the new product. Much work has been done
in identifying the analogous products and finding the parameter ranges for different products in different
industry sectors. Identifying a truly ‘analogous product’ has proven to be challenging. Many growth curve
implementations have shown that forecasting error could easily be more than 50% (Kahn, 2002; Goodwin
et al., 2013a). Hence the continued research in this area.
In predicting new product demand based on extracting the intention and expected behaviour of the cus-
tomers, various old and new techniques had been tried. The Delphi method, customer intention surveys,
and test marketing are some of the methods purely based on estimating the intentions directly. However,
some other techniques try to extract the same information indirectly and in a subtle way. Conjoint analysis
and choice models, agent-based models and prediction and preference markets are some of the successful
implementations in this strand.

3.2.8. Spare parts forecasting89

Spare parts are ubiquitous in modern societies. Their demand arises whenever a component fails or
requires replacement. Demand for spare parts is typically intermittent, which means that it can be fore-
casted using the plethora of parametric and non-parametric methods presented in section 2.7. In addition
to the intermittence of demand, spare parts have two additional characteristics that make them different
from Work-In-Progress and final products, namely: (i) they are generated by maintenance policies and part
breakdowns, and (ii) they are subject to obsolescence (Bacchetti and Saccani, 2012; Kennedy et al., 2002).
The majority of forecasting methods do not link the demand to the generating factors, which are often
related to maintenance activities. The demand for spare parts originates from the replacement of parts
in the installed base of machines (i.e., the location and number of products in use), either preventively or
upon breakdown of the part (Kim et al., 2017). Fortuin (1984) claims that using installed base information
to forecast the spare part demand can lead to stock reductions of up to 25%. An overview of the literature
that deals with spare parts forecasting with installed base information is given by Van der Auweraer and
Boute (2019). Spare parts demand can be driven by the result of maintenance inspections and, in this
case, a maintenance-based forecasting model should then be considered to deal with this issue. Such
forecasting models include the Delay Time (DT) model analysed in Wang and Syntetos (2011). Using the
fitted values of the distribution parameters of a data set related to a hospital pumps, Wang and Syntetos
(2011) have shown that when the failure and fault arriving characteristics of the items can be captured, it
is recommended to use the DT model to forecast the spare part demand with a higher forecast accuracy.
However, when such information is not available, then time series forecasting methods are recommended.

89 This subsection was written by Mohamed Zied Babai.

88
 Given the life cycle of products, spare parts are associated with a risk of obsolescence. Molenaers et al.
(2012) discussed a case study where 54% of the parts stocked at a large petrochemical company had seen
no demand for the last 5 years. Hinton (1999) reported that the US Department of Defence was holding
60% excess of spare parts, with 18% of the parts (with a total value of $1.5 billion) having no demand at all.
To take into account the issue of obsolescence in spare parts demand forecasting, Teunter et al. (2011) have
proposed the TSB method, which deals with linearly decreasing demand and sudden obsolescence cases.
By means of an empirical investigation based on the individual demand histories of 8000 spare parts SKUs
from the automotive industry and the Royal Air Force (RAF, UK), Babai et al. (2014) have demonstrated
the high forecast accuracy and inventory performance of the TSB method. Other variants of the Croston’s
method developed to deal with the risk of obsolescence in forecasting spare parts demand include the
Hyperbolic-Exponential Smoothing method proposed by Prestwich et al. (2014) and the modified Croston’s
method developed by Babai et al. (2019).

3.2.9. Predictive maintenance90

A common classification of industrial maintenance includes three types of maintenance (Montero Jimenez
et al., 2020). Corrective maintenance refers to maintenance actions that occur after the failure of a compo-
nent. Preventive maintenance consists of maintenance actions that are triggered after a scheduled number
of units as cycles, kilometers, flights, etc. To schedule the fixed time between two preventive maintenance
actions, the Weibull distribution is commonly used (Baptista et al., 2018). The drawbacks of preventive
maintenance are related to the replacement of components that still have a remaining useful life; there-
fore, early interventions imply a waste of resources and too late actions could imply catastrophic failures.
Additionally, the preventive intervention itself could be a source of failures too. Finally, predictive main-
tenance (PdM) complements the previous ones and, essentially, uses predictive tools to determine when
actions are necessary (Carvalho et al., 2019). Within this predictive maintenance group, other terms are
usually found in the literature as Condition-Based Maintenance and Prognostic and Health Management,
(Montero Jimenez et al., 2020).
The role of forecasting in industrial maintenance is of paramount importance. One application is to
forecast spare parts, whose demands are typically intermittent, usually required to carry out corrective
and preventive maintenances (Wang and Syntetos, 2011; Van der Auweraer and Boute, 2019). On the other
hand, it is crucial for PdM the forecast of the remaining useful time, which is the useful life left on an asset
at a particular time of operation (Si et al., 2011). This work will be focused on the latter, which is usually
found under the prognostic stage (Jardine et al., 2006).
The typology of forecasting techniques employed is very ample. Montero Jimenez et al. (2020) classify
them in three groups: physics-based models, knowledge-based models, and data-driven models. Physics-
based models require high skills on the underlying physics of the application. Knowledge-based models
are based on facts or cases collected over the years of operation and maintenance. Although, they are
useful for diagnostics and provide explicative results, its performance on prognostics is more limited.
In this sense, data-driven models are gaining popularity for the development of computational power,
data acquisition, and big data platforms. In this case, data coming from vibration analysis, lubricant
analysis, thermography, ultrasound, etc. are usually employed. Here, well-known forecasting models as

90 This subsection was written by Juan Ramón Trapero Arenas.

89
VARIMAX/GARCH are successfully used (Garcı́a et al., 2010; Cheng et al., 2012; Gomez Munoz et al.,
2014; Baptista et al., 2018). State Space models based on the Kalman Filter are also employed (Pedregal
and Carmen Carnero, 2006; Pedregal et al., 2009).
Recently, given the irruption of the Industry 4.0, physical and digital systems are getting more inte-
grated and Machine Learning/Artificial Intelligence are drawing the attention of practitioners and aca-
demics alike (Carvalho et al., 2019). In that same reference, it is found that the most frequently used
Machine Learning methods in PdM applications were Random Forest, Artificial Neural Networks, Support
Vector Machines and K-means.

3.2.10. Reverse logistics91

As logistics and supply chain operations rely upon accurate demand forecasts, reverse logistics and
closed loop supply chain operations rely upon accurate forecasts of returned items. Such items (usually
referred as cores) can be anything from reusable shipping or product containers to used laptops, mobile
phones or car engines. If some (re)manufacturing activity is involved in supply chains, it is both demand
and returned items forecasts that are needed since it is net demand requirements (demand – returns) that
drive remanufacturing operations.
Forecasting methods that are known to work well when applied to demand forecasting, such as SES for
example, do not perform well when applied to time-series of returns because they assume returns to be
a process independent of sales. There are some cases when this independence might hold, such as when
a recycler receives items sold by various companies and supply chains (Goltsos and Syntetos, 2020). In
these cases, simple methods like SES applied on the time series of returns might prove sufficient. Typically
though, returns are strongly correlated with past sales and the installed base (number of products with
customers). After all, there cannot be a product return if a product has not first been sold. This lagged
relationship between sales and returns is key to the effective characterisation of the returns process.
Despite the increasing importance of circular economy and research on closed loop supply chains,
returns forecasting has not received sufficient attention in the academic literature (notable contributions
in this area include Goh and Varaprasad, 1986; Toktay et al., 2000; Toktay, 2003; de Brito and van der
Laan, 2009; Clottey et al., 2012). The seminal work by Kelle and Silver (1989) offers a useful framework
to forecasting that is based on the degree of available information about the relationship between demand
and returns. Product level (PL) information consists of the time series of sales and returns, alongside
information on the time each product spends with a customer. The question then is how to derive this time
to return distribution. This can be done through managerial experience, by investigating the correlation of
the demand and the returns time series, or by serialising and tracking a subset (sample) of items. Past sales
can then be used in conjunction with this distribution to create forecasts of returns. Serial number level
(SL) information, is more detailed and consists of the time matching of an individual unit item’s issues and
returns and thus exactly the time each individual unit, on a serial number basis, spent with the customer.
Serialisation allows for a complete characterisation of the time to return distribution. Very importantly,
it also enables tracking exactly how many items previously sold remain with customers, providing time
series of unreturned past sales. Unreturned past sales can then be extrapolated — along with a time to
return distribution -– to create forecasts of returns.

91 This subsection was written by Aris A. Syntetos.

90
 Goltsos et al. (2019) offered empirical evidence in the area of returns forecasting by analysing a seri-
alised data set from a remanufacturing company in North Wales. They found the Beta probability distribu-
tion to best fit times-to-return. Their research suggests that serialisation is something worthwhile pursuing
for low volume products, especially if they are expensive. This makes a lot of sense from an investment
perspective, since the relevant serial numbers are very few. However, they also provided evidence that
such benefits expand in the case of high volume items. Importantly, the benefits of serialisation not only
enable the implementation of the more complex SL method, but also the accurate characterisation of the
returns process, thus also benefiting the PL method (which has been shown to be very robust).

3.3. Economics and finance

3.3.1. Macroeconomic survey expectations92
Macroeconomic survey expectations allow tests of theories of how agents form their expectations. Ex-
pectations play a central role in modern macroeconomic research (Gali, 2008). Survey expectations have
been used to test theories of expectations formation for the last 50 years. Initially the Livingston survey
data on inflationary expectations was used to test extrapolative or adaptive hypothesis, but the focus soon
turned to testing whether expectations are formed rationally (see Turnovsky and Wachter, 1972, for an
early contribution). According to (Muth, 1961, p.316), rational expectations is the hypothesis that: ‘expec-
tations, since they are informed predictions of future events, are essentially the same as the predictions of
the relevant economic theory.’ This assumes all agents have access to all relevant information. Instead, one
can test whether agents make efficient use of the information they possess. This is the notion of forecast
efficiency (Mincer and Zarnowitz, 1969), and can be tested by regressing the outturns on a constant and
the forecasts of those outturns. Under forecast efficiency, the constant should be zero and the coefficient
on the forecasts should be one. When the slope coefficient is not equal to one, the forecast errors will
be systematically related to information available at the forecast origin, namely, the forecasts, and cannot
be optimal. The exchange between Figlewski and Wachtel (1981, 1983) and Dietrich and Joines (1983)
clarifies the role of partial information in testing forecast efficiency (that is, full information is not neces-
sary), and shows that the use of the aggregate or consensus forecast in the individual realisation-forecast
regression outlined above will give rise to a slope parameter less than one when forecasters are efficient but
possess partial information. Zarnowitz (1985), Keane and Runkle (1990) and Bonham and Cohen (2001)
consider pooling across individuals in the realisation-forecast regression, and the role of correlated shocks
across individuals.
Recently, researchers considered why forecasters might not possess full-information, stressing infor-
mational rigidities: sticky information (see, inter alia, Mankiw and Reis, 2002; Mankiw et al., 2003), and
noisy information (see, inter alia, Woodford, 2002; Sims, 2003). Coibion and Gorodnichenko (2012, 2015)
test these models using aggregate quantities, such as mean errors and revisions.
Forecaster behaviour can be characterised by the response to new information. Over or under-reaction
would constitute inefficiency. Broer and Kohlhas (2018) and Bordalo et al. (2018) find that agents over-
react, generating a negative correlation between their forecast revision and error. The forecast is revised
by more than is warranted by the new information (over-confidence regarding the value of the new infor-
mation). Bordalo et al. (2018) explain the over-reaction with a model of ‘diagnostic’ expectations, whereas

92 This subsection was written by Michael P. Clements.

91
Fuhrer (2018) finds ‘intrinsic inflation persistence’: individuals under-react to new information, smoothing
their responses to news.
The empirical evidence is often equivocal, and might reflect: the vintage of data assumed for the out-
turns; whether allowance is made for ‘instabilities’ such as alternating over- and under-prediction (Rossi
and Sekhposyan, 2016) and the assumption of squared-error loss (see, for example, Patton and Timmer-
mann, 2007; Clements, 2014b).
Research has also focused on the histogram forecasts produced by a number of macro-surveys. Density
forecast evaluation techniques such as the probability integral transform93 have been applied to histogram
forecasts, and survey histograms have been compared to benchmark forecasts (see, for example, Bao et al.,
2007; Hall and Mitchell, 2009; Clements, 2018). Research has also considered uncertainty measures based
on the histograms Clements (2014a).
Engelberg et al. (2009) and Clements (2009, 2010) considered the consistency between the point predic-
tions and histogram forecasts. Reporting practices such as ‘rounding’ have also been considered (Binder,
2017; Manski and Molinari, 2010; Clements, 2011).
Clements (2019) reviews macroeconomic survey expectations.

3.3.2. Forecasting GDP and inflation94

As soon as Bayesian estimation of DSGEs became popular, these models have been employed in fore-
casting horseraces to predict the key macro variables, for example, Gross Domestic Product (GDP) and
inflation, as discussed in Del Negro and Schorfheide (2013). The forecasting performance is evaluated us-
ing rolling or recursive (expanded) prediction windows (for a discussion, see Cardani et al., 2015). DSGEs
are usually estimated using revised data, but several studies propose better results estimating the mod-
els using real-time data (see, for example, Del Negro and Schorfheide, 2013; Wolters, 2015; Kolasa and
Rubaszek, 2015b; Cardani et al., 2019).
The current DSGE model forecasting compares DSGE models to competitors. Among them, we can
include the Random Walk (the naive model which assumes a stochastic trend), the Bayesian VAR models
(Minnesota Prior à la Doan et al., 1984; and Large Bayesian VAR à la Bańbura et al., 2010), the Hybrid-
Models (the DSGE-VAR à la Del Negro and Schorfheide, 2004; and the DSGE-Factor Augmented VAR à
la Consolo et al., 2009), and the institutional forecasts (Greenbook, Survey Professional Forecasts, and the
Blue Chip, as illustrated in Edge and Gürkaynak, 2010).
Table 2 summarises the current DSGE forecasting literature mainly for the US and Euro Area and pro-
vided by estimating medium-scale models. As general findings, DSGEs can outperform other competitors,
with the exception for the Hybrid-Models, in the medium and long-run to forecast GDP and inflation. In
particular, Smets and Wouters (2007) was the first empirical evidence of how DSGEs can be competitive
with forecasts from Bayesian VARs, convincing researchers and policymakers in adopting DSGEs for pre-
diction evaluations. As discussed in Del Negro and Schorfheide (2013), the accuracy of DSGE forecasts
depends on how the model is able to capture low-frequency trends in the data. To explain the macro-
finance linkages during the Great Recession, the Smets and Wouters model was also compared to other

93 See, for example, Rosenblatt (1952), Shephard (1994), Kim et al. (1998), Diebold et al. (1998) and Berkowitz (2001).
94 This subsection was written by Alessia Paccagnini.

92
 Table 2: Alternative Competitors to DSGE Models

Competitor Reference
Hybrid Models US: Del Negro and Schorfheide (2004), Consolo et al. (2009)
Random Walk US: Gürkaynak et al. (2013), Euro Area: Warne et al. (2010),
Smets et al. (2014)
Bayesian VAR US: Smets and Wouters (2007), Gürkaynak et al. (2013), Wolters
(2015), Bekiros and Paccagnini (2014), Bekiros and Paccagnini
(2015a), Bekiros and Paccagnini (2015b), Euro Area: Warne et al.
(2010)
Time-Varying VAR and US: Bekiros et al. (2016), Euro Area: Bekiros and Paccagnini
Markov-Switching (2016)
Institutional Forecasts US: Edge and Gürkaynak (2010), Kolasa et al. (2012), Del Negro
and Schorfheide (2013), Wolters (2015)

DSGE specifications including the financial sector. For example, Del Negro and Schorfheide (2013), Ko-
lasa and Rubaszek (2015a), Galvão et al. (2016), and Cardani et al. (2019) provide forecasting performance
for DSGEs with financial frictions. This strand of the literature shows how this feature can improve the
baseline Smets and Wouters predictions for the business cycle, in particular during the recent Great Re-
cession.
However, the Hybrid-Models always outperform the DSGEs thanks to the combination of the theory-
based model (DSGE) and the statistical representation (VAR or Factor Augmented VAR), as illustrated by
Del Negro and Schorfheide (2004) and Consolo et al. (2009).
Moreover, several studies discuss how prediction performance could depend on the parameters’ esti-
mation. Kolasa and Rubaszek (2015b) suggest that updating DSGE model parameters only once a year is
enough to have accurate and efficient predictions about the main macro variables.

3.3.3. Forecasting UK unemployment95

Unemployment in the UK has fluctuated greatly over the last 160 years, ranging from near zero through-
out the two world wars, to over 18% during the UK’s Great Depression, with more than 3.5 million peo-
ple registered as unemployed; see figure 5. The data show many different ‘regimes’, both in levels and
volatility. Forecasting unemployment is challenging when the data are highly non-stationary with abrupt
distributional shifts, but persistence within regimes.
Two approaches dominate the literature on forecasting unemployment rates. The first is based on the
Phillips (1958) curve capturing a non-linear relationship between nominal wage inflation and the unem-
ployment rate, or the relation between unemployment and output described as Okun’s 1962 Law. The
second uses the time-series properties of the data to produce statistical forecasts, such as univariate lin-
ear models (e.g., ARIMA or unobserved component models), multivariate linear models (e.g., VARMA
or CVAR), various threshold autoregressive models (SETAR, TAR, STAR, LSTAR, etc.), Markov Switching
models and Artificial Neural Networks. The empirical literature is inconclusive as to the ‘best’ forecasting
models for unemployment.

95 This subsection was written by Jennifer L. Castle.

93
 0.175 UK unemployment rate

0.150

0.125

0.100

0.075

0.050

0.025

1860 1880 1900 1920 1940 1960 1980 2000 2020

Figure 5: Annual UK unemployment rate.

There is a huge literature forecasting the US unemployment rate, which is characterised by more regu-
lar business cycle patterns of expansion and contraction compared to the UK. Non-linear statistical models
tend to outperform within contractions or expansions, but perform worse across business cycles, see, e.g.,
Montgomery et al. (1998), Rothman (1998), and Koop and Potter (1999), whereas Proietti (2003) finds that
linear models characterised by higher persistence perform significantly better. For the UK, evidence of
non-linearities is found by Peel and Speight (2000), Milas and Rothman (2008), and Johnes (1999), and
Gil-Alana (2001) finds evidence of long-memory. Barnichon and Garda (2016) applies a flow approach to
unemployment forecasting and finds improvements, as does Smith (2011). See also the studies by Ace-
moglu and Scott (1994), Floros (2005), and Smith (2016).
One approach by Hendry (2001) assumes unemployment will fall when hiring is profitable. Profitabil-
ity is proxied by the gap between the real interest rate (reflecting costs) and the real growth rate (reflect-
ing the demand side), such that the unemployment rate rises when the real interest rate exceeds the real
growth rate, and vice versa. A dynamic equilibrium correction model, using impulse indicator saturation
(IIS: Hendry et al., 2008, and Johansen and Nielsen, 2009b) and step indicator saturation (SIS: Castle et al.,
2015c) to capture outliers and the different regimes forecasts accurately, also see Hendry (2015) and Castle
et al. (2020c).
Given the inconclusive state of the literature, how can we make progress in forecasting aggregate UK
unemployment? Forecasting models could be improved with either (i) better economic theories of aggre-
gate unemployment,96 or (ii) more general empirical models that tackle stochastic trends, breaks, dynam-
ics, non-linearities and interdependence,97 or better still, both. The COVID-19 pandemic and subsequent
lockdown policies highlight just how important forecasts of unemployment are, see Castle et al. (2020a).

96 There are many relevant theories based on microfoundations, including search and matching, loss of skills, efficiency wages, and
insider-outsider models, see Layard et al. (1991) for a summary. However, our focus is on aggregate unemployment.
97 See Hendry and Doornik (2014) for an approach to jointly tackling all of these issues.

94
Figure 6 shows forecasts from a model of unemployment which would have predicted a 1% rise in the
unemployment rate in April and May 2020 given a 30% decline in real GDP, based on the scenario of no
furlough scheme. Also included are extrapolative forecasts (Cardt: Doornik et al., 2020a) which show no
such rise, reflecting the success of the furlough scheme in maintaining employment during lockdown. As
these policies come to an end forecasting models are needed that can handle sudden shifts and changes in
policy.

1-step forecasts from unemployment model with ±2σ band

5.25 Unemployment rate April 2020 →
Cardt 1-step forecasts

5.00
↑
May
2020
4.75

4.50

4.25

4.00

3.75

2019 2020

Figure 6: 1-month ahead conditional forecasts for the unemployment rate for 2019(6)-2020(5)

3.3.4. Forecasting UK productivity98

The growth of labour productivity in the UK, measured by the percent change in output per hours
worked, has varied dramatically over the last 260 years ranging from -5.8% at the onset of the post WW1
economic crash to just over 7% in 1971; see panel A of figure 7. The data show that productivity growth
is very volatile and has undergone large historical shifts with productivity growth averaging around 1%
between 1800-1950 followed by an increase in the average annual growth to 3% between 1950-1975. Since
the mid-1970’s productivity growth has fluctuated around 2%; see panel B of figure 7. However, over the
last decade since 2009, 2% annual growth has proved to be an upper bound rather than the average.
The most common approach for forecasting productivity, both in the UK and elsewhere, is to esti-
mate the trend growth in productivity using aggregate data. For example, Gordon (2003) considers three
separate approaches for calculating trend labour productivity in the United States based on (i) average
historical growth rates outside of the business cycle, (ii) filtering the data using the Hodrick-Prescott (HP)
filter (Hodrick and Prescott, 1997), and (iii) filtering the data using the Kalman filter (Kalman, 1960). In
the UK, the Office for Budget Responsibility (OBR) follows a similar approach for generating its forecasts
of productivity based on the average historical growth rate as well as judgements about factors that may

98 This subsection was written by Andrew B. Martinez.

95
 Panel A. Historical Annual Growth Rates Panel B. Quarterly Year-over-year Growth Rates
8 8

6 6

4 4

2 2

0 0

-2 -2

-4 -4

-6 1775 1800 1825 1850 1875 1900 1925 1950 1975 2000 -6 1975 1980 1985 1990 1995 2000 2005 2010 2015 2020

Figure 7: UK Productivity (Output per total hours worked). Sources: Bank of England and Office for National Statistics.

cause productivity to deviate from its historical trend in the short-term.99 Alternative less commonly used
approaches include forecasting aggregate productivity using firm-level micro data (Bartelsman et al., 2011;
Bartelsman and Wolf, 2014) and using naive statistical models (Žmuk et al., 2018).
In the last few decades there have been several attempts to test and allow for time-varying trends in
productivity. However, the focus of these approaches has been primarily on the United States (Hansen,
2001; Roberts, 2001), which saw a sharp rise in productivity growth in the 1990’s that was not mirrored in
the UK (Basu et al., 2003). Tests for shifts in productivity growth rates in advanced economies did not find
evidence of a change in productivity growth in the UK until well after the financial crisis in 2007 (Benati,
2007; Turner and Boulhol, 2011; Glocker and Wegmüller, 2018).
A more recent approach by Martinez et al. (2019) allows for a time-varying long-run trend in UK pro-
ductivity. They are able to broadly replicate the OBR’s forecasts using a quasi-transformed autoregressive
model with one lag, a constant, and a trend. The estimated long-run trend is just over 2% per year through
2007 Q4 which is consistent with the OBR’s assumptions about the long-run growth rate of productivity
(OBR, 2019). However, it is possible to dramatically improve upon OBR’s forecasts in real-time by allow-
ing for the long-term trend forecast to adjust based on more recent historical patterns. By taking a local
average of the last four years of growth rates, Martinez et al. (2019) generate productivity forecasts whose
RMSE is on average 75% smaller than OBR’s forecasts extending five-years-ahead and is 84% smaller at
the longest forecast horizon.

3.3.5. Fiscal forecasting for government budget surveillance in Europe100

Recent economic recessions have led to a renewed interest in fiscal forecasting, mainly for deficit and
debt surveillance. This was certainly true in the case of the 2008 recession, and looks to become even more
important in the current economic crisis brought on by the COVID-19 pandemic. This is particularly im-
portant in Europe, where countries are subject to strong fiscal monitoring mechanisms. Two main themes

99 See https://obr.uk/forecasts-in-depth/the-economy-forecast/potential-output-and-the-output-gap. (Accessed:

2020-09-05)
100 This subsection was written by Diego J. Pedregal.

96
can be detected in the fiscal forecasting literature (Leal et al., 2008). First, investigate the properties of
forecasts in terms of bias, efficiency and accuracy. Second, check the adequacy of forecasting procedures.
The first topic has its own interest for long, mainly restricted to international institutions (Artis and
Marcellino, 2001). Part of the literature, however, argue that fiscal forecasts are politically biased, mainly
because there is usually no clear distinction between political targets and rigorous forecasts (Strauch et al.,
2004; Frankel and Schreger, 2013). In this sense, the availability of forecasts from independent sources
is of great value (Jonung and Larch, 2006). But it is not as easy as saying that independent forecasters
would improve forecasts due to the absence of political bias, because forecasting accuracy is compromised
by complexities of data, country-specific factors, outliers, changes in the definition of fiscal variables, etc.
Very often some of these issues are known by the staff of organisations in charge of making the official
statistics and forecasts long before the general public, and some information never leaves such institutions.
So this insider information is actually a valuable asset to improve forecasting accuracy (Leal et al., 2008).
As for the second issue, namely the accuracy of forecasting methods, the literature can be divided into
two parts, one based on macroeconomic models with specific fiscal modules that allows to analyse the
effects of fiscal policy on macro variables and vice versa (see Favero and Marcellino (2005) and references
therein), and the other based on pure forecasting methods and comparisons among them. This last stream
of research basically resembles closely what is seen in other forecasting areas: (i) there is no single method
outperforming the rest generally, (ii) judgemental forecasting is especially important due to data problems,
and (iii) combination of methods tends to outperform individual ones (Leal et al., 2008).
Part of the recent literature focused on the generation of very short-term public finance monitoring
systems using models that combine annual information with intra-annual fiscal data (Pedregal and Pérez,
2010). The idea is to produce global annual end-of-year forecasts of budgetary variables based on the most
frequently available fiscal indicators, so that changes throughout the year in the indicators can be used
as early warnings to infer the changes in the annual forecasts and deviations from fiscal targets (Pedregal
et al., 2014).
The level of disaggregation of the indicator variables are established according to the information avail-
able and the particular objectives. The simplest options are the accrual National Accounts annual or quar-
terly fiscal balances running on their cash monthly counterparts. A somewhat more complex version is the
previous one with all the variables broken down into revenues and expenditures. Other disaggregation
schemes have been applied, namely by region, by administrative level (regional, municipal, social secu-
rity, etc.), or by items within revenue and/or expenditure (VAT, income taxes, etc. Paredes et al., 2014;
Asimakopoulos et al., 2020).
Unfortunately, what is missing is a comprehensive and transparent forecasting system, independent of
Member States, capable of producing consistent forecasts over time and across countries. This is certainly
a challenge that no one has yet dared to take up.

3.3.6. Interest rate prediction with Markov switching models101

The (spot) rate on a (riskless) bond represents the ex-ante return (yield) to maturity which equates its
market price to a theoretical valuation. Interest rates and their term structure reflect two economic forces.
First, they depend on expected economic conditions over the life of the bonds; second, rates are affected

101 This subsection was written by Massimo Guidolin.

97
by the stance of monetary policy. Because business cycle conditions are characterised by recurring states
(like expansions and recessions) and monetary policy is subject to regimes, recent research has reported
that dynamic econometric models with regime shifts in parameters, and in particular Markov switching
models (MSMs), are useful at forecasting rates. This occurs not despite the non-linear nature of MSMs
but because of such nature (Ang and Bekaert, 2002b) that allows them to capture the pervasive of non-
linearities in the term structure (Stanton, 1997).
The usefulness of MS VAR models with term structure data had been established since Hamilton (1988)
and Garcia and Perron (1996): single-state, VARMA models are overwhelmingly rejected in favour of
multi-state models. Subsequently, a literature has emerged that has documented that MSMs are required
to successfully forecast the yield curve, with emphasis on short rates. Lanne and Saikkonen (2003) show
that a mixture of autoregressions with two regimes improves the predictions of US T-bill rates. Ang and
Bekaert (2002a) investigate the MS dynamics of short-term rates for the US, the UK, and Germany. Bekaert
et al. (2001) show that while single-state VARs cannot generate distributions consistent with the expecta-
tions hypothesis (EH), Markov regimes weaken the evidence against the EH. In fact, because the EH and
rational expectations imply that forward rates represent the optimal forecast of spot rates (apart from a
risk premium), Guidolin and Timmermann (2009) find that the optimal combinations of lagged short and
forward rates depend on regimes so that the EH holds only in such states. A number of papers have also
investigated the presence of regimes in the typical factors (level, slope, and convexity) that characterise
the no-arbitrage dynamics of the term structure, showing the predictive benefits of incorporating MS (see,
for example, Guidolin and Pedio, 2019; Hevia et al., 2015). As widely documented (Guidolin and Thorn-
ton, 2018), the predictable component in the mean rates is hardly significant. As a result, the random
walk tends to predict well. However, density forecasts reflect all moments and the models that capture
the dynamics of higher-order moments tend to perform best. MSMs appear at the forefront of a class of
non-linear models that produce accurate density predictions (Hong et al., 2004; Maheu and Yang, 2016).
Another literature has strived to fit rates not only under the physical measure, i.e., in time series, but to
predict rates when MS enters the pricing kernel, the fundamental pricing operator. A few papers have also
assumed that regimes represent a new risk factor (Dai and Singleton, 2003). This literature reports that
MSMs lead to a range of shapes for nominal and real term structures (Veronesi and Yared, 1999). Often the
model specifications that are not rejected by formal tests include MS (Ang et al., 2008; Bansal and Zhou,
2002). MSMs can reproduce the empirical puzzles that plague the EH (Bansal et al., 2004; Dai et al., 2007).
Finally, MSMs were proposed to forecast conditional variances. Cai (1994) developed a MS ARCH
model to examine volatility persistence because this may be inflated by regimes. Gray (1996) generalised
this attempt to MS GARCH and reported improvements in pseudo out-of-sample predictions. Further
advances in the methods and applications of MS GARCH are Haas et al. (2004) and Smith (2002).

3.3.7. Term structure interest rates prediction with threshold models102

Modelling and predicting default-free, short-term interest rates (also called ex-ante yields) are crucial
tasks in asset pricing and risk management. Indeed, the value of interest rate–sensitive securities depends
on the value of the riskless rate. Besides, the short interest rate is a fundamental ingredient in the for-
mulation and transmission of the monetary policy. However, many popular models of the short rate (for

102 This subsection was written by Manuela Pedio.

98
instance, continuous time, diffusion models) fail to deliver accurate out-of-sample forecasts. Their poor
predictive performance may depend on the fact that the stochastic behaviour of short interest rates may be
time-varying (for instance, it may depend on the business cycle and on the stance of monetary policy).
Notably, the presence of nonlinearities in the conditional mean and variance of the short-term yield
influences the behaviour of the entire term structure of spot rates implicit in riskless bond prices, the
relationship linking rates to their residual time-to-maturity. For instance, the level of the short-term rate
directly affects the slope of the yield curve. More generally, nonlinear rate dynamics imply a nonlinear
equilibrium relationship between short and the long-term yields.
A few studies have tried to capture the time-varying, nonlinear dynamics of interest rates using thresh-
old models. As discussed by Pai and Pedersen (1999) threshold models show a key advantage compared
their closest competitor, i.e., the Markov switching models: the regimes are not determined by an unob-
served latent variable, thus fostering interpretability. In most of the applications to interest rates, the
regimes are determined by the lagged level of the short rate itself, in a self-exciting fashion. For in-
stance, Pfann et al. (1996) explored nonlinear dynamics of the US short-term interest rate using a (self-
exciting) threshold autoregressive model augmented by conditional heteroscedasticity (namely, a TAR-
GARCH model) and found strong evidence of the presence of two regimes. More recently, also Gospodinov
(2005) has used a TAR-GARCH to predict the short-term rate and has shown that this model can capture
some well-documented features of the empirical data, such as high persistence and strong conditional
heteroscedasticity. Lemke and Archontakis (2008) have derived the analytical solution for arbitrage-free
bond yields when the short-term riskless rate follows a threshold autoregressive model and have found that
self-exciting autoregressive models are especially suitable to capture the near unit-root behaviour typically
observed in the evolution of short rates.
Other papers have tried to model the riskless yield curve using multivariate threshold models. For
instance, extending seminal work by Enders and Granger (1998), Clements and Galvão (2003) have tested
the expectation theory of the term structure of the interest rates in a nonlinear system that allows the
response of the change in short rates to past values of the spread between long and short term rates to
depend on the level of the spread. They have found that the spread predicts change in the short-term
interest rate only when its current level is high. As typical applications of the threshold regime switching
frameworks to empirical finance, a distinct literature exists that estimate TAR models to also forecast
conditional higher order moments, see, for example, Pfann et al. (1996) and more recently Dellaportas
et al. (2007), and all reported reasonable accuracy.
Despite their appeal and their good empirical performance, the use of threshold models to predict
interest rates has so far remained limited. This is mainly due to the fact that statistical inference for this
kind of regime switching models poses some challenges, because the likelihood function is discontinuous
with respect to the threshold parameters.

3.3.8. House price forecasting103

The boom and bust in housing markets in the early and mid 2000s and its decisive role in the Great
Recession has generated a vast interest in the dynamics of house prices and emphasised the importance
of accurately forecasting property price movements during turbulent times. International organisations,

103 This subsection was written by Alisa Yusupova.

99
central banks and research institutes have become increasingly engaged in monitoring the property price
developments across the world.104 At the same time, a substantial empirical literature has developed that
deals with predicting future house price movements (for a comprehensive survey see Ghysels et al., 2013).
Although this literature concentrates almost entirely on the US (see, for example, Rapach and Strauss, 2009;
Bork and Møller, 2015), there are many other countries, such as the UK, where house price forecastability
is of prime importance. Similarly to the US, in the UK, housing activities account for a large fraction of
GDP and of households’ expenditures; real estate property comprises a significant component of private
wealth and mortgage debt constitutes a main liability of households (Office for National Statistics, 2019).
The appropriate forecasting model has to reflect the dynamics of the specific real estate market and
take into account its particular characteristics. In the UK, for instance, there is a substantial empirical
literature that documents the existence of strong spatial linkages between regional markets, whereby the
house price shocks emanating from southern regions of the country, and in particular Greater London,
have a tendency to spread out and affect neighbouring regions with a time lag (see, for example, Cook and
Thomas, 2003; Holly et al., 2010; Antonakakis et al., 2018, inter alia).
Recent evidence also suggests that the relationship between real estate valuations and conditioning
macro and financial variables displayed a complex of time-varying patterns over the previous decades
(Aizenman and Jinjarak, 2013). Hence, predictive methods that do not allow for time-variation in both
predictors and their marginal effects may not be able to capture the complex house price dynamics in the
UK (see Yusupova et al., 2019, for a comparison of forecasting accuracy of a battery of static and dynamic
econometric methods).
An important recent trend is to attempt to incorporate information from novel data sources (such as
newspaper articles, social media, etc.) in forecasting models as a measure of expectations and percep-
tions of economic agents. It has been shown that changes in uncertainty about house prices impact on
housing investment and real estate construction decisions (Cunningham, 2006; Banks et al., 2015; Oh and
Yoon, 2020), and thus incorporating a measure of uncertainty in the forecasting model can improve the
forecastability of real estate prices. For instance in the UK, the House Price Uncertainty (HPU) index
(Yusupova et al., 2020), constructed using the methodology outlined in Baker et al. (2016),105 was found to
be important in predicting property price inflation ahead of the house price collapse of the third quarter of
2008 and during the bust phase (Yusupova et al., 2019). Along with capturing the two recent recessions (in
the early 1990s and middle 2000s) this index also reflects the uncertainly related to the EU Referendum,
Brexit negotiations and COVID-19 pandemic.

3.3.9. Exchange rate forecasting106

Exchange rates have long fascinated and puzzled researchers in international finance. The reason is
that following the seminal paper of Meese and Rogoff (1983), the common wisdom is that macroeconomic
models cannot outperform the random walk in exchange rate forecasting (see Rossi, 2013, for a survey).
This view is difficult to reconcile with the strong belief that exchange rates are driven by fundamentals,

104 For instance, the International Monetary Fund recently established the Global Housing Watch, the Globalisation and Monetary
Policy Institute of the Federal Reserve Bank of Dallas initiated a project on monitoring international property price dynamics, and
the UK Housing Observatory initiated a similar project for the UK national and regional housing markets.
105 For a comparison of alternative text-based measures of economic uncertainty see Kalamara et al. (2020).
106 This subsection was written by Michał Rubaszek.

100
such as relative productivity, external imbalances, terms of trade, fiscal policy or interest rate disparity
(MacDonald, 1998; Lee et al., 2013; Couharde et al., 2018). These two contradicting assertions by the
academic literature is referred to as “exchange rate disconnect puzzle”.
The literature provides several explanations for this puzzle. First, it can be related to the forecast
estimation error. The studies in which models are estimated with a large panels of data (Mark and Sul,
2001; Engel et al., 2008; Ince, 2014), long time series (Lothian and Taylor, 1996) or calibrated (Ca’ Zorzi and
Rubaszek, 2020) deliver positive results on exchange rate forecastability. Second, there is ample evidence
that the adjustment of exchange rates to equilibrium is non-linear (Taylor and Peel, 2000; Curran and
Velic, 2019), which might diminish the out-of-sample performance of macroeconomic models (Kilian and
Taylor, 2003; Lopez-Suarez and Rodriguez-Lopez, 2011). Third, few economists argue that the role of
macroeconomic fundamentals may be varying over time and this should be accounted for in a forecasting
setting (Byrne et al., 2016; Beckmann and Schussler, 2016).
The dominant part of the exchange rate forecasting literature investigates which macroeconomic model
performs best out-of-sample. The initial studies explored the role of monetary fundamentals to find that
these models deliver inaccurate short-term and not so bad long-term predictions in comparison to the
random walk (Meese and Rogoff, 1983; Mark, 1995). In a comprehensive study from mid-2000s, Cheung
et al. (2005) showed that neither monetary, uncovered interest parity (UIP) nor behavioural equilibrium
exchange rate (BEER) model are able to outperform the no-change forecast. A step forward was made by
Molodtsova and Papell (2009), who proposed a model combining the UIP and Taylor rule equations and
showed that it delivers competitive exchange rate forecasts. This result, however, has not been confirmed
by more recent studies (Cheung et al., 2019; Engel et al., 2019). In turn, Ca’ Zorzi and Rubaszek (2020)
argue that a simple method assuming gradual adjustment of the exchange rate towards the level implied
by the Purchasing Power Parity (PPP) performs well over shorter as well as longer horizon. This result
is consistent with the results of Ca’ Zorzi et al. (2017) and Eichenbaum et al. (2017), who showed that
exchange rates are predictable within a general equilibrium DSGE framework, which encompasses an
adjustment of the exchange rate to a PPP equilibrium. Finally, Ca’ Zorzi et al. (2020) discuss how extending
the PPP framework for other fundamentals within the BEER framework is not helping in exchange rate
forecasting. Overall, at the current juncture it might be claimed that “exchange rate disconnect puzzle”
is still puzzling, with some evidence that methods based on PPP and controlling the estimation forecast
error can deliver more accurate forecast than the random walk benchmark.

3.3.10. Financial time series forecasting with range-based volatility models107

The range-based (RB) volatility models is a general term for the models constructed with high and low
prices, and most often with their difference i.e., the price range. A short review and classification of such
models is contained in subsection 2.3.15. From practical point of view, it is important that low and high
prices are almost always available with daily closing prices for financial series. The price range (or its
logarithm) is a significantly more efficient estimator of volatility than the estimator based on closing prices
(Alizadeh et al., 2002). Similarly the co-range (the covariance based on price ranges) is a significantly
more efficient estimator of the covariance of returns than the estimator based on closing prices (Brunetti

107 This subsection was written by Piotr Fiszeder.

101
and Lildholdt, 2002). For these reasons models based on the price range and the co-range better describe
variances and covariances of financial returns than the ones based on closing prices.
Forecasts of volatility from simple models like moving average, EWMA, AR, ARMA based on the RB
variance estimators are more accurate than the forecasts from the same models based on squared returns
of closing prices (Vipul and Jacob, 2007; Rajvanshi, 2015). Forecasts of volatility from the AR model based
on the Parkinson estimator are more precise even than the forecasts from the standard GARCH models
based on closing prices (Li and Hong, 2011).
In plenty of studies it was shown that forecasts of volatility of financial returns from the univariate
RB models are more accurate than the forecasts from standard GARCH models based on closing prices
(see, for example, Mapa, 2003 for the GARCH-PARK-R model; Chou, 2005 for the CARR model; Fiszeder,
2005 for the GARCH-TR model; Brandt and Jones, 2006 for the REGARCH model; Chen et al., 2008 for
the TARR model; Lin et al., 2012 for the STARR model; Fiszeder and Perczak, 2016 for the GARCH model
estimated with low, high and closing prices during crisis periods; Molnár, 2016 for the RGARCH model).
The use of daily low and high prices in the multivariate volatility models leads to more accurate fore-
casts of covariance or covariance matrix of financial returns than the forecasts from the models based on
closing prices (see, for example, Chou et al., 2009 for the RB DCC model; Harris and Yilmaz, 2010 for
the hybrid EWMA model; Fiszeder, 2018 for the BEKK-HL model; Fiszeder and Fałdziński, 2019 for the
co-range DCC model; Fiszeder et al., 2019 for the DCC-RGARCH model).
The RB models were used in many financial applications. They lead for example to more precise fore-
casts of value-at-risk measures in comparison to the application of only closing prices (see, for example,
Chen et al., 2012 for the threshold CAViaR model; Asai and Brugal, 2013 for the HVAR model; Fiszeder
et al., 2019 for the DCC-RGARCH model; Meng and Taylor, 2020 for scoring functions). The application
of the multivariate RB models provides also the increase in the efficiency of hedging strategies (see, for
example, Chou et al., 2009 for the RB DCC model; Harris and Yilmaz, 2010 for the hybrid EWMA model;
Su and Wu, 2014 for the RB-MS-DCC model). Moreover, the RB volatility models have more significant
economic value than the return-based ones in the portfolio construction (Chou and Liu, 2010 for the RB
DCC model; Wu and Liang, 2011 for the RB-copula model). Some studies show that based on the fore-
casts from the volatility models with low and high prices it is possible to construct profitable investment
strategies (He et al., 2010 for the VECM model; Kumar, 2015 for the CARRS model).

3.3.11. Copula forecasting with multivariate dependent financial times series108

In this section, we focus on the practical advances on jointly forecasting multivariate financial time
series with copulas. In the copula framework, because marginal models and copula models are separa-
ble, point forecasts are straightforward with marginal models, but dependence information is ignored. A
joint probabilistic forecast with copulas involves both estimations of the copula distribution and marginal
models.
In financial time series, an emerging interest is to model and forecast the asymmetric dependence. A
typical asymmetric dependence phenomenon is that two stock returns exhibit greater correlation during
market downturns than market upturns. Patton (2006a) employs the asymmetric dependence between
exchange rates with a time-varying copula construction with AR and GARCH margins. A similar study for

108 This subsection was written by Feng Li.

102
measuring financial contagion with copulas allows the parameters of the copula to change with the states
of the variance to identify shifts in the dependence structure in times of crisis (Rodriguez, 2007).
In stock forecasting, Almeida and Czado (2012) employ a stochastic copula autoregressive model to
model DJI and Nasdaq, and the dependence at the time is modelled by a real-valued latent variable, which
corresponds to the Fisher transformation of Kendall’s τ. Li and Kang (2018) use a covariate-dependent
copula framework to forecast the time varying dependence that improves both the probabilistic forecasting
performance and the forecasting interpretability. Liquidity risk is another focus in finance. Weiß and
Supper (2013) forecast three types of liquidity-adjusted intraday Value-at-Risk (L-IVaR) with a vine copula
structure. The liquidity-adjusted intraday VaR is based on simulated portfolio values, and the results are
compared with the realised portfolio profits and losses.
In macroeconomic forecasting, most existing reduced-form models for multivariate time series produce
symmetric forecast densities. Gaussian copulas with skew Student’s-t margins depict asymmetries in the
predictive distributions of GDP growth and inflation (Smith and Vahey, 2016). Real-time macroeconomic
variables are forecasted with heteroscedastic inversion copulas (Smith and Maneesoonthorn, 2018) that
allow for asymmetry in the density forecasts, and both serial and cross-sectional dependence could be
captured by the copula function (Loaiza-Maya and Smith, 2020).
Copulas are also widely used to detect and forecast default correlation, which is a random variable
called time-until-default to denote the survival time of each defaultable entity or financial instrument (Li,
2000). Then copulas are used in modelling the dependent defaults (Li, 2000), forecasting credit risk (Bi-
elecki and Rutkowski, 2013), and credit derivatives market forecasting (Schönbucher, 2003). A much large
volume of literature is available for this specific area. See the aforementioned references therein. For par-
ticular applications in credit default swap (CDS) and default risk forecasting see Oh and Patton (2018) and
Li and He (2019) respectively.
In energy economics, Aloui et al. (2013) employ the time-varying copula approach, where the marginal
models are from ARMA(p,q)–GARCH(1,1) to investigate the conditional dependence between the Brent
crude oil price and stock markets in the Central and Eastern European transition economies. Bessa et al.
(2012) propose a time-adaptive quantile-copula where the copula density is estimated with a kernel density
forecast method. The method is applied to wind power probabilistic forecasting and shows its advantages
for both system operators and wind power producers. Vine copula models are also used to forecast wind
power farms’ uncertainty in power system operation scheduling. Wang et al. (2017) shows vine copulas
have advantages of providing reliable and sharp forecast intervals, especially in the case with limited
observations available.

3.3.12. Financial forecasting with neural networks109

Neural Networks (NNs) are capable of successfully modelling non-stationary and non-linear series.
This property has made them one of the most popular (if not the most) non-linear specification used by
practitioners and academics in Finance. For example, 89% of European banks use NNs to their operations
(European Banking Federation, 2019) while 25.4% of the NNs applications in total is in Finance (Wong
et al., 1995).

109 This subsection was written by Georgios Sermpinis.

103
 The first applications of NNs in Finance and currently the most widespread, is in financial trading.
In the mid-80s when computational power became cheaper and more accessible, hedge fund managers
started to experiment with NNs in trading. Their initial success led to even more practitioners to apply
NNs and nowadays 67% of hedge fund managers use NNs to generate trading ideas (BarclayHedge, 2018).
A broad measure of the success of NNs in financial trading is provided by the Eurekahedge AI Hedge Fund
Index110 where it is noteworthy the 13.02% annualised return of the selected AI hedge funds over the last
10 years.
In academia, financial trading with NNs is the focus of numerous papers. Notable applications of NNs
in trading financial series were provided by Kaastra and Boyd (1996), Tenti (1996), Panda and Narasimhan
(2007), Zhang and Ming (2008), and Dunis et al. (2010). The aim of these studies is to forecast the sign
or the return of financial trading series and based on these forecasts to generate profitable trading strate-
gies. A process that imitates how practitioners apply NNs in financial trading. The second major field
of applications of NNs in Finance is in derivatives pricing and financial risk management. The growth
of the financial industry and the provided financial services have made NNs and other machine learning
algorithms a necessity for tasks such as fraud detection, information extraction and credit risk assessment
(Buchanan, 2019). In derivatives pricing, NNs try to fill the limitations of the Black-Scholes model and
are being used in options pricing and hedging. In academia notable applications of NNs in risk manage-
ment are provided by Locarek-Junge and Prinzler (1998) and Liu (2005) and in derivatives by Bennell and
Sutcliffe (2004) and Psaradellis and Sermpinis (2016).
As discussed before, financial series due to their non-linear nature and their wide applications in prac-
tice seems the perfect forecasting data set for researchers that want to test their NN topologies. As a result,
there are thousands of forecasting papers in the field of NNs in financial forecasting. However, caution is
needed in interpretation of their results. NNs are sensitive to the choice of their hyperparameters. For a
simple MLP, a practitioner needs to set (among others) the number and type of inputs, the number of hid-
den nodes, the momentum, the learning rate, the number of epochs and the batch size. This complexity in
NN modelling leads inadvertently to the data snooping bias. In other words, a researcher that experiments
long enough with the parameters of a NN topology can have excellent in-sample and out-of-sample results
for a series. However, this does not mean that the results of his NN can be generalised. This issue has led
the related literature to be stained by studies cannot be extended in different samples.

3.3.13. Forecasting returns to investment style111

Investment style or factor portfolios are constructed from constituent securities on the basis of a variety
of a-priori observable characteristics, thought to affect future returns. For example a ‘Momentum’ portfo-
lio might be constructed with positive (‘long’) exposures to stocks with positive trailing 12-month returns,
and negative (‘short’) exposure to stocks with negative trailing 12-month returns (for full background and
context, see, for example Bernstein, 1995; Haugen, 2010).112 Explanations as to why such characteristics
seem to predict returns fall in to two main camps: firstly that the returns represent a risk premium, earned

110 https://www.eurekahedge.com/Indices/IndexView/Eurekahedge/683/Eurekahedge-AI-Hedge-fund-Index (Accessed: 2020-09-

01)
111 This subsection was written by Ross Hollyman.
112 The website of Kenneth French is an excellent source of data on investment style factor data and research.
http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data library.html

104
by the investor in return for taking on some kind of (undiversifiable) risk, and secondly that such returns
are the result of behavioural biases on the part of investors. In practice, both explanations are likely to
drive style returns to a greater or lesser extent. Several such strategies have generated reasonably con-
sistent positive risk-adjusted returns over many decades, but as with many financial return series, return
volatility is large relative to the mean, and there can be periods of months or even years when returns de-
viate significantly from their long-run averages. The idea of timing exposure to styles is therefore at least
superficially attractive, although the feasibility of doing so is a matter of some debate (Arnott et al., 2016;
Asness, 2016; Bender et al., 2018). Overconfidence in timing ability has a direct cost in terms of trading
frictions and opportunity cost in terms of potential expected returns and diversification forgone.
A number of authors write on the general topic of style timing (recent examples include Hodges et al.,
2017; Dichtl et al., 2019), and several forecasting methodologies have been suggested, falling in to three
main camps:

1. Serial Correlation: Perhaps the most promising approach is exploiting serial correlation in style re-
turns. Tarun and Bryan (2019) and Babu et al. (2020) outline two such approaches and Ehsani and
Linnainmaa (2020) explore the relationship between momentum in factor portfolios and momentum
in underlying stock returns. As with valuation spreads mentioned below, there is a risk that using
momentum signals to time exposure to momentum factor portfolios risks unwittingly compounding
exposure. A related strand of research relates (own) factor volatility to future returns, in particular
for momentum factors (Barroso, 2015; Daniel and Moskowitz, 2016).
2. Valuation Spreads: Using value signals (aggregated from individual stock value exposures) to time
exposure to various fundamental-based strategies is a popular and intuitively appealing approach
(Asness (2016)); however evidence of value added from doing so is mixed, and the technique seems
to compound risk exposure to value factors.
3. Economic & Financial Conditions: Polk et al. (2020) explore how economic and financial conditions
affect style returns (an idea that dates back at least to Bernstein (1995) and references therein).

Style returns exhibit distinctly non-normal distributions. On a univariate basis, most styles display
returns which are highly negatively skewed and demonstrate significant kurtosis. The long-run low corre-
lation between investment styles is often put forward as a benefit of style-based strategies, but more careful
analysis reveals that non-normality extends to the co-movements of investment style returns; factors ex-
hibit significant tail dependence. Christoffersen and Langlois (2013) explores this issue, also giving details
of the skew and kurtosis of weekly style returns. These features of the data mean that focusing solely on
forecasting the mean may not be sufficient, and building distributional forecasts becomes important for
proper risk management. Jondeau (2007) writes extensively on modelling non-gaussian distributions.

3.3.14. Forecasting stock returns113

Theory and intuition suggest a plethora of potentially relevant predictors of stock returns – the era of
big data further increases the data available for forecasting returns. When forecasting with large numbers
of predictors, conventional ordinary least squares (OLS) estimation is highly susceptible to overfitting,

113 This subsection was written by David E. Rapach.

105
which is exacerbated by the substantial noise in stock return data (reflecting the intrinsically large unpre-
dictable component in returns).
Over the last decade or so, researchers have explored methods for forecasting returns with large num-
bers of predictors. Principal component regression extracts the first few principal components (or factors)
from the set of predictors; the factors then serve as predictors in a low-dimensional predictive regression,
which is estimated via OLS. Intuitively, the factors combine the information in the individual predictors
to reduce the dimension of the regression, which helps to guard against overfitting. Ludvigson and Ng
(2007) find that a few factors extracted from hundreds of macroeconomic and financial variables improve
out-of-sample forecasts of the US market return. Kelly and Pruitt (2013) and Huang et al. (2015b) use par-
tial least squares (Wold, 1966) to construct target-relevant factors from a cross section of valuation ratios
and a variety of sentiment measures, respectively, to improve market return forecasts.
Since Bates and Granger (1969), it has been known that combinations of individual forecasts often per-
form better than the individual forecasts themselves (Timmermann, 2006). Rapach et al. (2010) show that
forecast combination can significantly improve out-of-sample market return forecasts. They first construct
return forecasts via individual univariate predictive regressions based on numerous popular predictors
from the literature (Goyal and Welch, 2008). They then compute a simple combination forecast by taking
the average of the individual forecasts. Rapach et al. (2013) demonstrate that forecast combination exerts
a strong shrinkage effect, thereby helping to guard against overfitting.
An emerging literature uses machine-learning techniques to construct forecasts of stock returns based
on large sets of predictors. In an investigation of lead-lag relationships among developed equity markets,
Rapach et al. (2013) appear to be the first to employ machine-learning tools to predict market returns. They
use the elastic net (ENet, Zou and Hastie, 2005), a generalisation of the popular least absolute shrinkage
and selection operator (LASSO, Tibshirani, 1996). The LASSO and ENet employ penalised regression
to guard against overfitting in high-dimensional settings by shrinking the parameter estimates toward
zero. Chinco et al. (2019) use the LASSO to forecast high-frequency (one-minute-ahead) individual stock
returns and report improvements in out-of-sample fit, while Rapach et al. (2019) use the LASSO to improve
monthly forecasts of industry returns.
Incorporating insights from Diebold and Shin (2019), Han et al. (2020) use the LASSO to form com-
bination forecasts of cross-sectional stock returns based on a large number of firm characteristics from
the cross-sectional literature (e.g., Harvey et al., 2016; McLean and Pontiff, 2016; Hou et al., 2020), ex-
tending the conventional OLS approach of Haugen and Baker (1996), Lewellen (2015), and Green et al.
(2017). Rapach and Zhou (2020) and Dong et al. (2020) use the LASSO and ENet to compute combination
forecasts of the market return based on popular predictors from the time-series literature and numerous
anomalies from the cross-sectional literature, respectively. Forecasting individual stock returns on the
basis of firm characteristics in a panel framework, Freyberger et al. (2020) and Gu et al. (2020) employ
machine-learning techniques – such as the nonparametric additive LASSO (Huang et al., 2010), random
forests (Breiman, 2001), and artificial neural networks – that allow for nonlinear predictive relationships.

106
3.3.15. Forecasting crashes in stock markets114
Time series data on financial asset returns have special features. Returns themselves are hard to fore-
cast, while it seems that volatility of returns can be predicted. Empirical distributions of asset returns show
occasional clusters of large positive and large negative returns. Large negative returns, that is, crashes
seem to occur more frequently than large positive returns. Forecasting upcoming increases or decreases in
volatility can be achieved by using variants of the Autoregressive Conditional Heteroskedasticity (ARCH)
model (Engle, 1982; Bollerslev, 1986) or realized volatility models (Taylor, 1986). These models take (func-
tions of) past volatility and past returns as volatility predictors, although also other explanatory variables
can be incorporated in the regression.
An important challenge that remains is to predict crashes. Sornette (2003) summarises potential causes
for crashes and these are computer trading, increased trading in derivatives, illiquidity, trade and budget
deficits, and especially, herding behaviour of investors. Yet, forecasting the exact timing of crashes may
seem impossible, but on the other hand, it may be possible to forecast the probability that a crash may
occur within a foreseeable future. Given the herding behaviour, any model to use for prediction should
include some self-exciting behaviour. For that purpose, Aı̈t-Sahalia et al. (2015) propose mutually exciting
jump processes, where jumps (read: crashes as negative jumps) can excite new jumps, also across assets
or markets (see also Chavez-Demoulin et al., 2005). Another successful approach is the Autoregressive
Conditional Duration (ACD) model (Engle and Russell, 1997, 1998), which refers to a time series model
for durations between (negative) events.
An alternative view on returns’ volatility and the potential occurrence of crashes draws upon the earth-
quake literature (Ogata, 1978, 1988). The idea is that tensions in and across tectonic plates build up, until
an eruption, and after that, tension starts to build up again until the next eruption. By modelling the
tension-building-up process using so-called Hawkes processes (Hawkes, 1971, 2018; Hawkes and Oakes,
1974; Ozaki, 1979), one can exploit the similarities between earthquakes and financial crashes. Gresnigt
et al. (2015) take Hawkes processes to daily S&P 500 data and show that it is possible to create reliable
probability predictions of a crash occurrence within the next five days. Gresnigt et al. (2017a,b) further
develop a specification strategy for any type of asset returns, and document that there are spillovers across
assets and markets.
Given investor behaviour, past crashes can ignite future crashes. Hawkes processes are particularly
useful to describe this feature and can usefully be implemented to predict the probability of nearby crashes.
By the way, these processes can also be useful to predict social conflicts, as also there one may discern
earthquake-like patterns. van den Hengel and Franses (2020) document their forecasting power for social
conflicts in Africa.

3.4. Energy
3.4.1. Building energy consumption forecasting115
In Europe, buildings account for 40% of total energy consumed and 36% of total CO2 emissions (Patti
et al., 2016). Given that energy consumption of buildings is expected to increase in the coming years, fore-
casting their electricity consumption becomes critical for improving energy management and planning.

114 This subsection was written by Philip Hans Franses.

115 This subsection was written by Evangelos Spiliotis.

107
 The main challenge in energy consumption forecasting is that building energy systems are complex
in nature, with their behaviour depending on various factors related to the type (e.g., residential, office,
entertainment, business, and industrial) and the end-uses (e.g., heating, cooling, hot water, and lighting)
of the building, its construction, its occupancy, the occupants’ behaviour and schedule, the efficiency of
the installed equipment, and the weather conditions (Zhao and Magoulès, 2012). Special events, holidays,
and calendar effects can also affect the behaviour of the systems and further complicate the consumption
patterns, especially when forecasting at hourly or daily level. As a result, producing accurate forecasts
typically requires developing customised, building-specific methods.
To deal with this task, the literature focuses on three main classes of forecasting methods, namely en-
gineering, statistical, and ML (Mat Daut et al., 2017). Engineering methods (typically utilised through
software tools such as DOE-2, EnergyPlus, BLAST, and ESP-r) build on physical models that forecast con-
sumption through detailed equations which account for the particularities of the building (Al-Homoud,
2001; Zhao and Magoulès, 2012). Although engineering methods can provide accurate forecasts, they re-
quire high level of expertise and computational resources, do not easily adapt to pattern changes, and
depend on detailed information which may be difficult to collect or even assume (Foucquier et al., 2013).
Statistical models (mainly linear regression, ARIMA, ARIMAX, and exponential smoothing) are fast to
compute, require less information as input, can easily adapt to pattern changes, and be effectively ad-
justed for different types of buildings (Deb et al., 2017). Yet, statistical methods prescribe the behaviour
of the systems, while also ignoring building characteristics, thus often leading to inaccurate results. Fi-
nally, ML methods (mainly neural networks, support vector machines, decision trees and fussy models)
are data-driven and more generic, having also the capacity to account for multiple non-linear dependen-
cies between the energy consumed and the factors influencing its value (Ahmad et al., 2014). However,
data, computational, and expertise limitations may become relevant (Makridakis et al., 2020c). Drawing
from the above, the literature has been inconclusive about which class of methods is the most appropriate,
with the conclusions drawn being subject to the examined building type, dataset used, forecasting horizon
considered, and data frequency at which the forecasts are produced (Wei et al., 2019). To mitigate this
problem, combinations of methods and hybrids have been proposed, reporting encouraging results (Zhao
and Magoulès, 2012; Mohandes et al., 2019).
Other practical issues refer to data pre-processing. Energy consumption data is typically collected at
high frequencies through smart meters and therefore display noise and missing or extreme values due to
monitoring issues. As a result, verifying the quality of the input data through diagnostics and data cleans-
ing techniques, as well as optimising the selected time frames, are important for improving forecasting
performance (Bourdeau et al., 2019). Similarly, it is critical to select appropriate regressor variables which
are of high quality and possible to accurately predict to assist energy consumption forecasting. Finally, it
must be carefully decided whether the bottom-up, the top-down or a combination method will be used for
producing forecasts at both building and end-use level (Kuster et al., 2017; Spiliotis et al., 2020c).

3.4.2. Forecasting for building energy optimisation116

Building energy optimisation is an important topic nowadays as it is expected to play a significant role
on our way to a carbon-free future. An example for a typical application in this space is optimisation of

116 This subsection was written by Christoph Bergmeir.

108
the heating, ventilation, and air conditioning (HVAC) system. The goal is to minimise its energy use under
the constraints of maintaining certain comfort levels in the building. Though this is predominantly an
optimisation exercise, forecasting comes in at different points of the system as input into the optimisation,
and many problems in this space involve forecasting as a sub-problem, for example: energy consumption
forecasting, room occupancy forecasting, inside temperature forecasting, (hyper-local) forecasts of outside
temperature, air pressure forecasting for ventilation, and others.
For example, Krüger and Givoni (2004) use a linear regression approach to predict inside temperatures
in 3 houses in Brazil and optimally select the thermal systems that should be used based on their partic-
ularities, and Ruano et al. (2006) propose the use of a neural network to predict temperatures in a school
building and optimally control air-conditioned systems. Madaus et al. (2020) predict hyper-local extreme
heat events, combining global climate models and machine learning models. Jing et al. (2018) predict air
pressure to tackle the air balancing problem in ventilation systems with a support vector machine. Predict-
ing energy demand on a building/household level from smart meter data is an important research topic not
only for energy savings. In the building space, Ahmad et al. (2017), Touzani et al. (2018), and Wang et al.
(2018b) predict building energy consumption of residential and commercial buildings to improve energy
efficiency and achieve energy savings. They use decision tree-based algorithms (random forests, gradient
boosted trees), and neural networks for their predictions. Also, recent trends in forecasting such as global
forecasting models built across a set of time series (Januschowski et al., 2020), especially using (recurrent)
neural networks (Bandara et al., 2020a; Hewamalage et al., 2020) are particularly suitable for this type of
processing due to their capabilities to deal with external inputs and cold-start problems. Such capabilities
are necessary if there are different regimes in the simulations under which to predict, an example of such
a system for HVAC optimisation is presented by Godahewa et al. (2020).
More generally, many challenges in the space of building energy optimisation are classical examples
of so-called “predict then optimise” problems (Demirovic et al., 2019; Elmachtoub and Grigas, 2017).
Here, different possible scenario predictions are obtained from different assumptions in the form of input
parameters. These input parameters are then optimised to achieve a desired predicted outcome. As both
prediction and optimisation are difficult problems, they are usually treated separately (Elmachtoub and
Grigas, 2017), though there are now recent works where they are considered together (Elmachtoub and
Grigas, 2017; El Balghiti et al., 2019; Demirovic et al., 2019), and this will certainly be an interesting
avenue for future research.

3.4.3. Electricity price forecasting117

Forecasting electricity prices has various challenges that are highlighted in the detailed review paper
by Weron (2014). Even though there are economically well motivated fundamental electricity price mod-
els, forecasting models based on evaluating historic price data are the dominating the academic literature.
In recent years the focus on probabilistic forecasting grew rapidly, as they are highly relevant for many ap-
plications in energy trading and risk management, storage optimization and predictive maintenance, (Ziel
and Steinert, 2018; Nowotarski and Weron, 2018). Electricity price data is highly complex and is influenced
by regulation. However, there is electricity trading based on auctions and on continuous trading. Many
markets like the US and European markets organize day-ahead auctions for electricity prices, see figure 8.

117 This subsection was written by Florian Ziel.

109
Thus, we have to predict multivariate time series type data, (Ziel and Weron, 2018). In contrast, intraday
markets usually apply continuous trading to manage short term variations due to changes in forecasts of
renewable energy and demand, and outages (Kiesel and Paraschiv, 2017).
The key challenge in electricity price forecasting is to address all potential characteristics of the con-
sidered market, most notably (some of them visible in figure 8):

1. (time-varying) autoregressive effects and (in)stationarity

2. calendar effects (daily, weekly and annual seasonality, holiday effects, clock-change)
3. impacts from external inputs, that partially have to be predicted in advance: load/demand/consumption,
power generation (especially from wind and solar), relevant fuel prices (e.g., oil, coal, natural gas,
emission allowances), related power market prices (future, balancing and neighboring markets),
availabilities of power plants and interconnectors, import/export related data, and weather data
4. (time-varying) volatility and higher moment effects
5. price spikes (positive and negative)
6. price clustering

In recent years, statistical and machine learning methods gained a lot of attraction in day-ahead elec-
tricity price forecasting. Even though the majority of effects is linear there are specific non-linear de-
pendencies that can be explored by using non-linear models, especially neural networks (Dudek, 2016;
Lago et al., 2018; Ugurlu et al., 2018; Marcjasz et al., 2019). Of course this comes along with higher com-
putational costs compared to linear models. Fezzi and Mosetti (2020) illustrate that even simple linear
models can give highly accurate forecasts, if correctly calibrated. However, there seems to be consensus
that forecast combination is appropriate, particularly for models that have different structures or different
calibration window length (Gaillard et al., 2016; Mirakyan et al., 2017; Hubicka et al., 2018).
Another increasing stream of electricity price forecasting models does not focus on the electricity price
itself, but the sale/supply and purchase/demand curves of the underlying auctions (see figure 8, but also
Ziel and Steinert, 2016; Shah and Lisi, 2020; Mestre et al., 2020). This sophisticated forecasting problem
allows more insights for trading applications and the capturing of price clusters.
In forecasting intraday markets the literature just started to grow quickly. As the aforementioned mar-
ket characteristics get less distinct if information from day-ahead markets is taken into account appropri-
ately. However, intraday prices are usually more volatile and exhibit more stronger price spikes. Thus,
probabilistic forecasting is even more relevant (Janke and Steinke, 2019; Narajewski and Ziel, 2020b). Re-
cent studies showed that European markets are close to weak-form efficiency. Thus naive point forecasting
benchmarks perform remarkably well (Oksuz and Ugurlu, 2019; Narajewski and Ziel, 2020a; Marcjasz
et al., 2020).

3.4.4. Hybrid machine learning system for short-term load forecasting118

Short-term load forecasting (STLF) horizon spans from 1 to 168 hours ahead (Hong and Fan, 2016). It
is regarded as the basis for power system operation. Processes like unit commitment and optimal power
flow rely on STLF (Saksornchai et al., 2005; Bo and Li, 2012). Additionally, since competitive energy
markets continually evolve, STLF becomes vital for new market entities such as retailers, aggregators and

118 This subsection was written by Ioannis Panapakidis.

110
 80 Supply curve 80
Tue Wed Thu Fri Sat Sun Mon Tue Wed Thu Fri Sat Sun Mon Demand curve
Intersection line
60 60

40 40
Electricity Price [EUR/MWh]

Electricity Price [EUR/MWh]

20 20

pgridc
0 0

−20 −20

Import Biomass
−40 Pump storage Waste −40
Nuclear Wind onshore
Lignite Wind offshore
−60 Hard coal Solar −60
Gas Load
Oil Electricity price
−80 Hydro power −80

60 600

Bid Volume [MW]

400
xtime[idxsub] curvs
dts[csub, dtx]

40
Power [GW]

200
0
20
−200
0 −400
−600
−20
12May 14May 16May 18May 20May 22May 24May 20 25 30 35
Time (2020−05−12 to 2020−05−25) Aggregated Volume [GW]
curvs[dtx]

Figure 8: Hourly German day-ahead electricity price data resulting from a two-sided auction (top left) with corresponding 24
sale/supply and purchase/demand curves for 24 May 2020 and highlighted curves for 17:00 (top right), power generation and con-
sumption time series (bottom left), and bid structure of 24 May 2020 17:00 (bottom right).

prosumers for applications such as strategic bidding, portfolio optimisation and tariff design (Danti and
Magnani, 2017; Ahmad et al., 2019). Due to the importance of STLF, a vast number of researchers have
proposed algorithms of different complexity and structure during the last years (Hahn et al., 2009).
Towards increasing the prediction accuracy of a single algorithm in STLF studies, hybrid models have
been developed and tested. The term “hybrid” refers to the synergy of two or more algorithms. For more
details, see section 2.6.12, but also the works of Bozkurt et al. (2017), López et al. (2017), Lu et al. (2019),
and El-Hendawi and Wang (2020). Below, we provide an indicative list of successful implementations of
hybrid systems on STLF.
Hybrid systems has been tested on validation data (through forecasting competitions), power system
aggregated load, and application oriented tasks. Ma (2021) proposed an ensemble method based on a com-
bination of various single forecasters on GEFCom2012 forecasting competition data that outperformed
benchmark forecasters such as Theta method, NN, ARIMA, and others. For aggregated load cases, re-
111
searchers focus on different countries and energy markets. Zhang et al. (2018) combined empirical mode
decomposition (EMD), ARIMA, and wavelet neural networks (WNN) optimised by the fruit fly algorithm
on Australian Market data and New York City data. Their approach was to separate the linear and non-
linear components from original electricity load; ARIMA is used for linear part while the WNN for the
non-linear one. Singh and Dwivedi (2019) proposed a multi-objective optimisation for the proper calibra-
tion of an NN parameters. The data set involves data from ISO New England and ERCOT markets.
Sideratos et al. (2020) proposed that a radial basis network that performs the initial forecasting could
serve as input to a convolutional neural network that performs the final forecasting. The STLF application
involved two systems, i.e., Crete Island and Hellenic Interconnected System. In all cases, the proposed
model led to lower error compared to the persistence model, NN, and SVM. The application oriented tasks
involved small scale load. Semero et al. (2020) focused on the energy management of a microgrid located in
China using EMD to decompose the load, adaptive neuro-fuzzy inference system (ANFIS) for forecasting
and particle swarm intelligence (PSO) to optimize ANFIS parameters. The results show that the proposed
approach yielded superior performance over four other methods. Faraji et al. (2020) proposed a hybrid
system for the scheduling of a prosumer microgrid in Iran. Various machine learning algorithms provided
load and weather forecasts. Through an optimisation routine, the best individual forecast is selected. The
hybrid system displayed better accuracy from the sole application of the individual forecasters. Xie et al.
(2020) focused on the load of two rooms in building located in China. The authors proposed an operation
strategy of a ground source heat pump system integrated with cooling storage system through the volume
control of stored water. The strategy is based on the hour-ahead load forecasting provided by LSTM-BPNN
model. Compared to conventional control, the load forecasting based control results in reductions of the
energy consumption and operating cost.

3.4.5. Predicting spikes in energy markets119

Deregulated electricity markets have been studied in the last two decades focusing on demand and
prices. Several stylised facts of prices, which are crucial for predictive purposes, have been identified:
seasonality, mean reversion, volatility, and spikes (or jumps). Very high prices on electricity markets are
usually observed in connection to a sudden increase in demand during extreme weather events or as a
consequence of drops in the supply in case of plant failures. The presence of spikes is explored in several
papers which can be divided into two main streams: spike forecasting and prediction of prices under
normal regime through robust estimators.
Within the first set of papers, the work by Lu et al. (2005) tries to explain the origins of price spikes em-
ploying a new composite index and develops a model based on data mining to forecast price spikes taking
advantage of the links with a set of explanatory variables. Spikes are often modelled as one regime of non-
linear models for time series. This approach is followed by Mount et al. (2006) and Becker et al. (2008). The
first paper focuses on regime-switching models with parameters driven by time-varying variables. These
models prove very effective in predicting price spikes thanks to the strict connection of transition prob-
abilities with current market conditions. The second paper adopts Markov switching models for spikes
prediction including demand and weather variables for the estimation of time-varying transition proba-
bilities. Christensen et al. (2009, 2012) suggest treating and forecasting price spikes through Poisson au-

119 This subsection was written by Luigi Grossi.

112
toregressive and discrete-time processes, respectively. In the first paper, spikes are modelled to account for
their persistence, while in the second a non-linear variant of the autoregressive conditional hazard model
is used to exploit the information of previous spikes. Herrera and González (2014) use a Hawkes model
combined with extreme events theory to forecast short term spikes in electricity markets which are then
applied to improve one-day-ahead Value at Risk (VaR) forecasting. Interregional links among different
electricity markets are used by Clements et al. (2015) and Manner et al. (2016) to forecast electricity price
spikes. A new procedure for the simulation of electricity spikes has been recently proposed by Muniain
and Ziel (2020).
The second stream of literature includes papers developing outlier detection methods or robust esti-
mators to improve the forecasting performance of the models. Martı́nez-Álvarez et al. (2011) tackle the
issue of outlier detection using the concept of motifs, defined as patches of units preceding observations
marked as anomalous in a training set. The prediction of outliers is then used to improve the general fore-
casting performance of the model. Janczura et al. (2013) focus on the detection and treatment of outliers
in electricity price to improve the general forecasting performance of time series models based on accurate
estimation of seasonal and stochastic components, both using original and filtered data. A very similar
approach, based on seasonal autoregressive models and outlier filtering, is followed by Afanasyev and Fe-
dorova (2019). Recently, Grossi and Nan (2019) introduced a procedure for the robust statistical prediction
of electricity prices. The econometric framework is represented by the non-linear SETAR process and the
robustification is obtained by the application of GM estimators relying on different weighting functions.
The focus of the paper is not on the prediction of spikes, but on the forecast of ordinary prices using esti-
mators which are slightly affected by the presence of abnormal observations. A similar approach has been
followed by Wang et al. (2020a) using an outlier-robust machine learning algorithm.

3.4.6. Crude oil price forecasting120

Crude oil, one of the leading energy resources, has contributed to over one-third of the world’s energy
consumption (Alvarez-Ramirez et al., 2003). The fluctuations of the crude oil price have a significant
impact on industries, governments as well as individuals, with substantial up-and-downs of the crude oil
price bringing dramatic uncertainty for the economic and political development (Kaboudan, 2001; Cunado
and De Gracia, 2005). Thus, it is critical to develop reliable methods to accurately forecast crude oil price
movement, so as to guard against the crude oil market extreme risks and improve macroeconomic policy
responses. However, the crude oil price movement suffers from complex features such as nonlinearity,
irregularities, dynamics and high volatility (Kang et al., 2009; Alquist et al., 2013; Herrera et al., 2018),
making the crude oil price forecasting still one of the most challenging forecasting problems.
Some prior studies have suggested that the crude oil price movement is inherently unpredictable, and
it would be pointless and futile to attempt to forecast future prices, see Miao et al. (2017) for a detailed
summary. These agnostics consider the naive no-change forecast as the best available forecast value of
future prices. In recent years, however, numerous studies result in forecasts that are more accurate than
naive no-change forecasts, making the forecasting activities of crude oil prices promising (Alquist et al.,
2013; Baumeister et al., 2015).

120 This subsection was written by Xiaoqian Wang.

113
 Extensive research on crude oil price forecasting has focused predominantly on the econometric mod-
els, such as VAR, ARCH-type, ARIMA, and Markov models (see, for example, Mirmirani and Li, 2004; Ag-
nolucci, 2009; Mohammadi and Su, 2010; e Silva et al., 2010). In the forecasting literature, unit root tests
are commonly applied to examine the stationarity of crude oil prices prior to econometric modelling (Silva-
pulle and Moosa, 1999; Serletis and Rangel-Ruiz, 2004; Rahman and Serletis, 2012). It is well-documented
that crude oil prices are driven by a large set of external components, which are themselves hard to pre-
dict, including supply and demand forces, stock market activities, oil-related events (e.g., war, weather
conditions), political factors, etc. In this context, researchers have frequently considered structural mod-
els, which relate the oil price movements to a set of economic factors. With so many econometric models,
is there an optimal one? Recently, de Albuquerquemello et al. (2018) proposed a SETAR model, allowing
for predictive regimes changing after a detected threshold, and achieved performance improvements over
six widely used econometric models. Despite their high computational efficiency, the econometric models
are generally limited in the ability to nonlinear time series modelling.
On the other hand, artificial intelligence and machine learning techniques, such as belief networks,
support vector machines (SVMs), recurrent neural networks (RNNs), and extreme gradient boosting (XG-
Boost), provided powerful solutions to recognise the nonlinear and irregular patterns of the crude oil price
movement with high automation (see, for example, Abramson and Finizza, 1991; Xie et al., 2006; Ming-
ming and Jinliang, 2012; Gumus and Kiran, 2017). However, challenges also exist in these techniques, such
as computational cost and overfitting. In addition, a large number of studies have increasingly focused on
the hybrid forecasting models based on econometrics models and machine learning techniques (Jammazi
and Aloui, 2012; He et al., 2012; Baumeister and Kilian, 2015; Chiroma et al., 2015), achieving improved
performance. Notably, the vast majority of the literature has focused primarily on the deterministic pre-
diction, with much less attention paid to the probabilistic prediction and uncertainty analysis. However,
the high volatility of crude oil prices makes probabilistic prediction more crucial to reduce the risk in
decision-making (Abramson and Finizza, 1995; Sun et al., 2018).

3.4.7. Forecasting renewable energy technologies121

The widespread adoption of renewable energy technologies (RETs) plays a driving role in the transi-
tion to low-carbon energy systems, a key challenge to face climate change and energy security problems.
Forecasting the diffusion of RETs is critical for planning a suitable energy agenda and setting achievable
targets in terms of electricity generation, although the available time series are often very short and pose
difficulties in modelling. Among the approaches employed to predict the spread of RETs, the innovation
diffusion framework appears a suitable choice, for its ability to capture in a simple and efficient way the
nonlinear, evolutionary nature of these processes. From the seminal work by Marchetti and Nakicenovic
(1979), theorising that innovation dynamics in energy systems are similar to those of other commercial sec-
tors or technologies, a growing body of literature has treated energy technologies with innovation diffusion
models, as if they were new products needing to be accepted in a market on the basis of final decisions of
individuals and market mechanisms like incentive measures, tariffs and price changes. Examples of these
applications are provided in the papers by Guseo et al. (2007), Guidolin and Mortarino (2010), Dalla Valle
and Furlan (2014), Guidolin and Guseo (2012), and Furlan et al. (2016).

121 This subsection was written by Mariangela Guidolin.

114
 The basic rationale for applying this class of models to the energy context is that learning from oth-
ers’ experience, spread of relevant information and information sharing within communities are primary
drivers of these processes, giving rise to imitation dynamics well captured by diffusion models, like the
Bass model (Bass, 1969). On the other hand, the adoption of RETs is characterised by several technologi-
cal, financial and administrative complexities with high initial costs, that may act as barriers to individual
adoption. In a review on the application of diffusion models to renewables by Rao and Kishore (2010), it
has been noticed that renewables’ typical characteristics such as low load factor, need for energy storage,
small size, high upfront costs create a competitive disadvantage, while Meade and Islam (2015b) suggested
that renewable technologies are different from other industrial technological innovations because, in the
absence of focused support, they are not convenient from a financial point of view. In this sense, policy
measures and incentive mechanisms, such as feed-in tariffs, have been used to stimulate the market. In or-
der to capture the real effect of these measures, the Generalised Bass model (GMB) by Bass et al. (1994) has
proven to be an essential modelling tool, allowing to explain a substantial lack of pioneering consumers in
RETs adoption and the positive effect exerted by incentives in driving the market (see, in particular, Bunea
et al., 2020). The use of the GBM also allows a more accurate forecasting procedure through an efficient
description of the nonlinear trajectory, which, in the case of renewables is often far from being smooth.
A more recent stream of research has focused on forecasting the diffusion of renewables in a competi-
tive environment, accounting for the dominant role of non-renewable energy technologies. This involves
the use of multivariate models. See, for instance, the works by Guidolin and Guseo (2016), Furlan and
Mortarino (2018), and Guidolin and Alpcan (2019).

3.4.8. Wind power forecasting122

Wind energy is a leading source of renewable energy, meeting 4.8% of global electricity demand in
2018, more than twice that of solar energy (IEA, Paris, 2020). Kinetic energy in the wind is converted into
electrical energy by wind turbines according to a characteristic ‘power curve’. Power production is propor-
tion to the cube of the wind speed at low-to-moderate speeds, and above this is constant at the turbine’s
rated power. At very high or low wind speeds no power is generated. Furthermore, the power curve is
influenced by additional factors including air density, icing, and degradation of the turbine’s blades.
Forecasts of wind energy production are required from minutes to days-ahead to inform the operation
of wind farms, participation in energy markets and power systems operations. However, the limited pre-
dictability of the weather and the complexity of the power curve make this challenging. For this reason,
probabilistic forecasts are increasingly used in practice (Bessa et al., 2017). Their value for energy trading
is clear (Pinson et al., 2007), but quantifying value for power system operation is extremely complex. Wind
power forecasting may be considered a mature technology as many competing commercial offerings exist,
but research and development efforts to produce novel and enhanced products are ongoing.
Short-term forecasts (hours to days ahead) of wind power production are generally produced by com-
bining numerical weather predictions (NWP) with a model of the wind turbine, farm or even regional
power curve, depending on the objective. The power curve may be modelled using physical information,
e.g., provided by the turbine manufacturer, in which case it is also necessary to post-process NWP wind

122 This subsection was written by Jethro Browell.

115
speeds to match the same height-above-ground as the turbine’s rotor. More accurate forecasts can be pro-
duced by learning the NWP-to-energy relationship from historic data when it is available. State-of-the-art
methods for producing wind power forecasts leverage large quantities of NWP data to produce a single
forecast (Andrade et al., 2017) and detailed information about the target wind farm (Gilbert et al., 2020a).
A number of practical aspects may also need to be considered by users, such as maintenance outages and
requirements to reduce output for other reasons, such as noise control or electricity network issues.
Very short-term forecast (minutes to a few hours ahead) are also of value, and on these time scales recent
observations are the most significant input to forecasting models and more relevant than NWP. Classical
time series methods perform well, and those which are able to capture spatial dependency between mul-
tiple wind farms are state-of-the-art, notably vector autoregressive models and variants (Cavalcante et al.,
2016; Messner and Pinson, 2018). Care must be taken when implementing these models as wind power
time series are bounded by zero and the wind farm’s rated power meaning that errors may not be assumed
to be normally distributed. The use of transformations is recommended, though the choice of transforma-
tion depends on the nature of individual time series (Pinson, 2012).
Research is ongoing in a range of directions including: improving accuracy and reducing uncertainty
in short-term forecasting, extending forecast horizons to weeks and months ahead, and improving very
short-term forecast with remote sensing and data sharing (Sweeney et al., 2019).

3.4.9. Wave forecasting123

Ocean waves are primarily generated by persistent winds in one direction. The energy thus propagated
by the wind is referred to as wave energy flux and follows a linear function of wave height squared and
wave period. Wave height is typically measured as significant wave height, the average height of the highest
third of the waves. The mean wave period, typically measured in seconds, is the average time between the
arrival of consecutive crests, whereas the peak wave period is the wave period at which the highest energy
occurs at a specific point.
The benefit of wave energy is that it requires significantly less reserve compared to those from wind
and solar renewable energy sources (Hong et al., 2016). For example, the forecast error at one hour ahead
for the simulated wave farms is typically in the range of 5–7%, while the forecast errors for solar and wind
are 17 and 22% respectively (Reikard et al., 2011). Solar power is dominated by diurnal and annual cycles
but also exhibits nonlinear variability due to factors such as cloud cover, temperature and precipitation.
Wind power is dominated by large ramp events such as irregular transitions between states of high and
low power. Wave energy exhibits annual cycles and is generally smoother although there are still some
large transitions, particularly during the winter months. In the first few hours of forecasting wave energy,
time series models are known to be more accurate than numerical wave prediction. Beyond these forecast
horizons, numerical wave prediction models such as SWAN (Simulating WAves Nearshore, Booij et al.,
1999) and WAVEWATCH III® (Tolman, 2008) are widely used. As there is as yet no consensus on the most
efficient model for harnessing wave energy, potential wave energy is primarily measured with energy flux,
but the wave energy harnessed typically follows non-linear functions of wave height and wave period in
the observations of the six different types of wave energy converters (Reikard et al., 2015).

123 This subsection was written by Jooyoung Jeon.

116
 To model the dependencies of wind speed, wave height, wave period and their lags, Reikard et al.
(2011) uses linear regressions, which were then converted to forecasts of energy flux. Pinson et al. (2012)
uses Reikard et al.’s (2011) regression model and log-normal distribution assumptions to produce proba-
bilistic forecasts. López-Ruiz et al. (2016) model the temporal dependencies of significant wave heights,
peak wave periods and mean wave direction using a vector autoregressive model, and used them to pro-
duce medium to long term wave energy forecasts. Jeon and Taylor (2016) model the temporal dependencies
of significant wave heights and peak wave periods using a bivariate VARMA-GARCH to convert the two
probabilistic forecasts into a probabilistic forecast of wave energy flux, finding this approach worked bet-
ter than either univariate modelling of wave energy flux or bivariate modelling of wave energy flux and
wind speed. Taylor and Jeon (2018) produce probabilistic forecasts for wave heights using a bivariate
VARMA-GARCH model of wave heights and wind speeds, and using forecasts so as to optimise decision
making for scheduling offshore wind farm maintenance vessels dispatched under stochastic uncertainty.
On the same subject, Gilbert et al. (2020b) use statistical post-processing of numerical wave predictions to
produce probabilistic forecasts of wave heights, wave periods and wave direction and a logistic regression
to determine the regime of the variables. They further applied the Gaussian copula to model temporal
dependency but this did not improve their probabilistic forecasts of wave heights and periods.

3.4.10. Power forecasting from PV systems124

Over the past few years, a number of forecasting techniques for PV power systems has been developed
and presented in the literature. In general, the quantitative comparison among different forecast tech-
niques is challenging, as the factors influencing the performance are numerous: the historical data, the
weather forecast, the temporal horizon and resolution, and the installation conditions. A recent review
by Sobri et al. (2018) presents a comparative analysis of previous works, also including statistical errors.
However, since the conditions and metrics used in each work were different, the comparison is not very
meaningful. Dolara et al. (2018) present relevant evaluation metrics for PV forecasting accuracy, while
Leva et al. (2019) compare their effectiveness and immediate comprehension. In term of forecast horizon
for PV power systems, intraday (Nespoli et al., 2019) and the 24 hours of the next day (Mellit et al., 2020)
are considered the most important.
Nespoli et al. (2019) compared two of the most widely used and effective methods for the forecasting of
the PV production: a method based on Multi-Layer Perceptron (MLP) and a hybrid method using artificial
neural network combined with clear sky solar radiation. In the second case, the simulations are based on
a feed-forward neural network (FFNN) but, among the inputs, the irradiation in clear sky conditions is
provided. This method is called Physical Hybrid Artificial Neural Network (PHANN) and is graphically
depicted in figure 9 (Dolara et al., 2015). PHANN method demonstrates better performance than classical
NN methods. Figure 10 shows a comparison between the measured and forecasted hourly output power of
the PV plant for both sunny and cloudy days. The PHANN method shows good forecasting performance,
especially for sunny days.
Ogliari et al. (2017) compared the PV output power day-ahead forecasts performed by deterministic
(based on three and five parameters electric equivalent circuit) and stochastic hybrid (based on artificial
neural network models) methods aiming to find the best performance conditions. In general, there is

124 This subsection was written by Sonia Leva.

117
 𝐼𝑛𝑝𝑢𝑡 𝐿𝑎𝑦𝑒𝑟 𝐻𝑖𝑑𝑑𝑒𝑛 𝐿𝑎𝑦𝑒𝑟𝑠 𝑂𝑢𝑡𝑝𝑢𝑡 𝐿𝑎𝑦𝑒𝑟

Weather Forecast

𝑃𝐹

𝐶𝑙𝑒𝑎𝑟 𝑆𝑘𝑦 𝑀𝑜𝑑𝑒𝑙

𝐿𝑜𝑛𝑔, 𝐿𝑎𝑡, 𝑇𝑖𝑙𝑡, 𝐴𝑧𝑖𝑚𝑢𝑡ℎ

Figure 9: Physical Hybrid Artificial Neural Network (PHANN) for PV power forecasting.

Figure 10: Measured versus forecasted output power by MLP and PHANN methods.

no significant difference between the two deterministic models, with the three-parameter approach being
slightly more accurate. Figure 11 shows the daily value of normalised mean absolute error (NMAE%) for
216 days evaluated by using PHANN and three parameters electric circuit. The PHANN hybrid method
achieves the best forecasting results, and only a few days of training can provide accurate forecasts.
Dolara et al. (2018) analysed the effect of different approaches in the composition of a training data-set
for the day-ahead forecasting of PV power production based on NN. In particular, the influence of different
data-set compositions on the forecast outcome has been investigated by increasing the size of the training
set size and by varying the lengths of the training and validation sets, in order to assess the most effective
training method of this machine learning approach. As a general comment on the reported results, it can
be stated that a method that employs the same chronologically consecutive samples for training is best
suited when the availability of historical data is limited (for example, in newly deployed PV plant), while
training based on randomly mixed samples method, appears to be most effective in the case of a greater
data availability. Generally speaking, ensembles composed of independent trials are most effective.

118
 30 30
PHANN 10TrD PHANN 215TrD
3 PARAM 3 PARAM

25 25

20 20
NMAE %

NMAE %
15 15

10 10

5 5

0 0
20 40 60 80 100 120 140 160 180 200 20 40 60 80 100 120 140 160 180 200
day day

Figure 11: Daily NMAE% of the PHANN method trained with 10 days (left) and with 215 days (right) compared with the three-
parameters model.

3.4.11. Forecasting with simulation in energy planning125

In electrical power systems with renewable energy dependence, the power generators need to be sched-
uled to supply the system demand (de Queiroz, 2016). In general, for modeling renewables future be-
haviour, such as hydro, wind and solar photovoltaics (PV), stochastic scenarios should be included in the
scheduling, usually in a dispatch optimisation problem. Due to the complexity and uncertainly associated,
this problem is, in general, modeled with time series scenarios and multi-stage stochastic approaches.
de Queiroz (2016) presented a review for hydrothermal systems, with a focus on the optimisation algo-
rithms. Vargas et al. (2019), through a systematic literature network analysis, pointed out the main fore-
casting methods for wind power generation. For a PV solar power forecasting and optimisation review, see
Ahmed et al. (2020).
This subsection aims to emphasise the importance of forecasting with simulation in renewable energy
planning, especially in hydroelectric systems. In this context, due to the data spatial and temporal depen-
dence structure, time series models are useful for future scenarios generation. Although the proposal could
be forecasting for short-term planning and scheduling, simulation strategies are explored for considering
and estimating uncertainty in medium and/or long-term horizons.
According to Hipel and McLeod (1994), stochastic processes of natural phenomena, such as the re-
newables ones, are, in general, stationary. One of the main features of hydroelectric generation systems
is the strong dependence on hydrological regimes. To deal with this task, the literature focuses on two
main classes for forecasting/simulation streamflow data: physical and data-driven models (Zhang et al.,
2015). Water resources management for hydropower generation and energy planning is one of the main
challenges for decision-makers. At large, the hydrological data are transformed into the so-called afflu-
ent natural energy, that is used for scenarios simulation and serve as input for the optimisation algorithms
(Oliveira et al., 2015). The current state-of-the-art models for this proposal are the periodic ones. Hipel and
McLeod (1994) presented a wide range of possibilities, but the univariate periodic autoregressive (PAR) is
still the benchmark, with several enhanced versions. The approach fits a model to each period of the his-

125 This subsection was written by Fernando Luiz Cyrino Oliveira.

119
torical data and the residuals are simulated to generate new future versions of the time series, considered
stationary. Among many others, important variations and alternative proposals to PAR with bootstrap pro-
cedures, Bayesian dynamic linear models, spatial information and copulas versions are detailed in Souza
et al. (2012), Marangon Lima et al. (2014), Lohmann et al. (2016) and de Almeida Pereira and Veiga (2019),
respectively.
It is worth considering the need for renewables portfolio simulation. This led Pinheiro Neto et al.
(2017) to propose a model to integrate hydro, wind and solar power scenarios for Brazilian data. For
the Eastern United States, Shahriari and Blumsack (2018) add to the literature on the wind, solar and
blended portfolios over several spatial and temporal scales. For China, Liu et al. (2020) proposed a multi-
variable model, with a unified framework, to simulate wind and PV scenarios to compensate hydropower
generation. The key challenges and trends for forecasting with simulation in a portfolio of renewable
energy planning are still related to the inclusion of exogenous variables, such as climate, meteorological,
calendar and economic ones.

3.4.12. Collaborative forecasting in the energy sector126

As mentioned in section 3.4.8, the combination of geographically distributed time series data, in a col-
laborative forecasting (or data sharing) framework, can deliver significant improvements in the forecasting
accuracy of each individual renewable energy power plant. The same is valid for hierarchical load fore-
casting (Hong et al., 2019) and energy price forecasting (see section 3.4.3). However, this data might have
different owners, which are unwilling to share their data due to the following reasons: (i) personal or busi-
ness sensitive information, (ii) lack of understanding about which data can and cannot be shared, and (iii)
lack of information about economic (and technical) benefits from data sharing.
In order to tackle these limitations, recent research in energy time series forecasting is exploring two
alternative (and potentially complementary) pathways: (i) privacy-preserving analytics, and (ii) data mar-
kets.
The role of privacy-preserving techniques applied collaborative forecasting is to combine time series
data from multiple data owners in order to improve forecasting accuracy and keep data private at the
same time. For solar energy forecasting, Berdugo et al. (2011) described a method based on local and
global analog-search that uses solar power time series from neighbouring sites, where only the times-
tamps and normalised weights (based on similarity) are exchanged and not the time series data. Zhang
and Wang (2018) proposed, for wind energy forecasting with spatia-temporal data, a combination of ridge
linear quantile regression and Alternating Direction Method of Multipliers (ADMM) that enables each data
owner to autonomously solve its forecasting problem, while collaborating with the others to improve fore-
casting accuracy. However, as demonstrated by Gonçalves et al. (2020a), the mathematical properties of
these algorithms should be carefully analysed in order to avoid privacy breaches (i.e., when a third party
recovers the original data without consent).
An alternative approach is to design a market (or auction) mechanism for time series or forecasting
data where the data owners are willing to sell their private (or confidential) data in exchange for an eco-
nomic compensation (Agarwal et al., 2019). The basic concept consists in pricing data as a function of
privacy loss, but it can be also pricing data as a function of tangible benefits such as electricity market

126 This subsection was written by Ricardo Bessa.

120
profit maximization. Gonçalves et al. (2020b) adapted for renewable energy forecasting the model de-
scribed in Agarwal et al. (2019), by considering the temporal nature of the data and relating data price
with the extra revenue obtained in the electricity market due to forecasting accuracy improvement. The
results showed a benefit in terms of higher revenue resulting from the combination of electricity and data
markets. With the advent of peer-to-peer energy markets at the domestic consumer level (Parag and Sova-
cool, 2016), smart meter data exchange between peers is also expected to increase and enable collaborative
forecasting schemes. For this scenario, Yassine et al. (2015) proposed a game theory mechanism where a
energy consumer maximizes its reward by sharing consumption data and a data aggregator can this data
with a data analyst (which seeks data with the lowest possible price).
Finally, promoting data sharing via privacy-preserving or data monetisation can also solve data scarcity
problems in some use cases of the energy sector, such as forecasting the condition of electrical grid as-
sets (Fan et al., 2020). Moreover, combination of heterogeneous data sources (e.g., numerical, textual,
categorical) is a challenging and promising avenue of future research in collaborative forecasting (Obst
et al., 2019).

3.5. Environmental applications

3.5.1. Forecasting two aspects of climate change127
First into the Industrial Revolution, the UK is one of the first out: in 2013 its per capita CO2 emissions
dropped below their 1860 level, despite per capita real incomes being around 7-fold higher (Hendry, 2020).
The model for forecasting UK CO2 emissions was selected from annual data 1860-2011 on CO2 emissions,
coal & oil usage, capital and GDP, their lags and non-linearities. Figures 12(a) to 12(c) show the non-
stationary time series with strong upward then downward trends, punctuated by large outliers from world
wars, miners strikes plus shifts from legislation and technological change: Castle and Hendry (2020a).
Saturation estimation at 0.1% using Autometrics (Doornik, 2018) retaining all other regressors, detected 4
step shifts coinciding with major policy interventions like the 2008 Climate Change Act, plus numerous
outliers, revealing a cointegrated relation. The multi-step forecasts over 2012—2017 from a VAR in panel
(d) of figure 12 show the advantage of using step-indicator saturation (SIS: Castle et al., 2015c).
We formulated a 3-equation simultaneous model of atmospheric CO2 and Antarctic Temperature and
Ice volume over 800,000 years of Ice Ages in 1000-year frequency (Paillard, 2001; Kaufmann and Juselius,
2013). Driven by non-linear functions of eccentricity, obliquity, and precession (see panels (a), (b), and (c)
of figure 13 respectively), the model was selected with saturation estimation. Earth’s orbital path is calcu-
lable into the future (Croll, 1875 and Milankovitch, 1969), allowing 100,000 years of multi-step forecasts
at endogenous emissions. Humanity has affected climate since 10 thousand years ago (kya: Ruddiman,
2005), so we commence forecasts there. Forecasts over −10 to 100 with time series from 400kya in panels
(d) to (f) of figure 13 show paths within the ranges of past data ±2.2SE (Pretis and Kaufmann, 2018).
Atmospheric CO2 already exceeds 400ppm (parts per million), dramatically outside the Ice-Age range
(Sundquist and Keeling, 2009). Consequently, we conditionally forecast the next 100,000 years, simulating
the potential climate for anthropogenic CO2 (Castle and Hendry, 2020b) noting the ‘greenhouse’ tem-
perature is proportional to the logarithm of CO2 (Arrhenius, 1896). The orbital drivers will continue to

127 This section was written by David F. Hendry.

121
CO2 emissions, tons per capita →

12.5 225
(a) (b)
200
175
10.0 150
125
100
Coal (Mt)
7.5 75 Oil (Mt)
Natural Gas (Mtoe)
50
↑ Wind+Solar (Mtoe)
1860 2013→ 25
5.0 0
1875 1900 1925 1950 1975 2000 1875 1900 1925 1950 1975 2000
8
log(GDP) 600 (d)
log(Capital stock) (c)
7 500

6 400

5 300 Dynamic forecasts, SIS, RMSFE=29

Dynamic forecasts, no SIS, RMSFE=94
1900 1950 2000 2000 2005 2010 2015
Figure 12: UK emissions, energy sources, economic variables and multi-step forecasts.

Eccentricity Obliquity
(a) (b)
0.04 24

0.02 23

-400 -300 -200 -100 0 100 -400 -300 -200 -100 0 100
Precession Ice Forecasts Fitted
(c) (d)
0.3 5
0.2
4
0.1

-400 -300 -200 -100 0 100 -400 -300 -200 -100 0 100
CO2 Forecasts Fitted Temp Forecasts Fitted
300 5
(e) (f)
0
250
-5
200
-10

-400 -300 -200 -100 0 100 -400 -300 -200 -100 0 100
Figure 13: Ice-Age data and forecasts.

122
influence all three variables but that relation is switched off in the scenario for ‘exogenised’ CO2 . The 110
dynamic forecasts conditional on 400ppm and 560ppm with ±2SE bands are shown in figure 14, panels (a)
and (b) for Ice and Temperature respectively. The resulting global temperature rises inferred from these
Antarctic temperatures would be dangerous, at more than 5◦ C, with Antarctic temperatures positive for
thousands of years (Vaks et al., 2019; Pretis and Kaufmann, 2020).

Ice Forecasts at CO2=560ppm Temp Forecasts at CO2=400ppm

Fitted Forecasts at CO2=400ppm 20 Fitted Forecasts at CO2=560ppm
5
(a) 15 (b)
4
10
Peak Ice-Age temperature
3
↑ 5 ↓
2
Least Ice-Age ice volume
0

1 -5
-10
0
-400 -300 -200 -100 0 100 -400 -300 -200 -100 0 100
Figure 14: Ice-Age simulations.

3.5.2. Weather forecasting128

The weather has a huge impact on our lives, affecting health, transport, agriculture, energy use, and
leisure. Since Bjerknes (1904) introduced hydrodynamics and thermodynamics into meteorology, weather
prediction has been based on merging physical principles and observational information. Modern weather
forecasting is based on numerical weather prediction (NWP) models that rely on accurate estimates of the
current state of the climate system, including ocean, atmosphere and land surface. Uncertainty in these
estimates is propagated through the NWP model by running the model for an ensemble of perturbed initial
states, creating a weather forecast ensemble (Buizza, 2018; Toth and Buizza, 2019).
One principal concern in NWP modelling is that small-scale phenomena such as clouds and convec-
tive precipitation are on too small a scale to be represented directly in the models and must, instead, be
represented by approximations known as parameterisations. Current NWP model development aims at
improving both the grid resolution and the observational information that enters the models (Bannister
et al., 2020; Leuenberger et al., 2020). However, for fixed computational resources, there is a trade-off
between grid resolution and ensemble size, with a larger ensemble generally providing a better estimate of
the prediction uncertainty. Recent advances furthermore include machine learning approaches to directly
model the small-scale processes, in particular cloud processes (see, for example, Gentine et al., 2018; Rasp
et al., 2018).
Despite rapid progress in NWP modelling, the raw ensemble forecasts exhibit systematic errors in
both magnitude and spread (Buizza, 2018). Statistical post-processing is thus routinely used to correct
systematic errors in calibration and accuracy before a weather forecast is issued, see Vannitsem et al. (2018)
for a recent review. A fundamental challenge here is to preserve physical consistency across space, time
and variables (see, for example, Möller et al., 2013; Schefzik et al., 2013; Heinrich et al., 2020). This is

128 This subsection was written by Thordis Thorarinsdottir.

123
particularly important when the weather forecast is used as input for further prediction modelling, e.g., in
hydrometeorology (Hemri et al., 2015; Hemri, 2018).
At time scales beyond two weeks, the weather noise that arises from the growth of the initial uncer-
tainty, becomes large (Royer, 1993). Sources of long-range predictability are usually associated with the
existence of slowly evolving components of the earth system, including the El Niño Southern Oscillation
(ENSO), monsoon rains, the Madden Julian Oscillation (MJO), the Indian Ocean dipole, and the North At-
lantic Oscillation (NAO), spanning a wide range of time scales from months to decades (Vitart et al., 2012;
Hoskins, 2013). It is expected that, if a forecasting system is capable of reproducing these slowly evolving
components, they may also be able to forecast them (Van Schaeybroeck and Vannitsem, 2018). The next
step is then to find relationships between modes of low-frequency variability and the information needed
by forecast users such as predictions of surface temperature and precipitation (Roulin and Vannitsem,
2019; Smith et al., 2020).

3.5.3. Air quality forecasting129

To preserve human health, European Commission stated in the Directive (2008/50/EC) that member
states have to promptly inform the population when the particulate matter (PM) daily mean value exceeds
(or is expected to exceed) the threshold of 50µg/m3 . Therefore, systems have been designed in order to
produce forecasts for up to three days in advance using as input the measured value of concentration and
meteorological conditions. These systems can be classified in (i) data-driven models (Carnevale et al., 2016;
Stadlober et al., 2018; Corani, 2005), and (ii) deterministic chemical and transport models (Honoré et al.,
2007; Manders et al., 2009). In this section, a brief overview of the application of these systems to the high
polluted area of Lombardy region, in Italy, will be presented.
Carnevale et al. (2018) compared the results of three different forecasting systems based on neural
networks, lazy learning models, and regression trees respectively. A single model has been identified for
each monitoring station. In the initial configuration, only the last three PM measurements available were
used to produce the forecast. In this configuration, the systems offered reasonable performance, with
correlation coefficients ranging from 0.6 (lazy learning method) to 0.75 (neural network). The work also
demonstrated that the performance of the ensemble of the three systems was better than the best model
for each monitoring station.
Starting from the results of this work, a second configuration was implemented, using as input also the
wind speed measured in the meteorological monitoring station closest to the measurement point of PM.
The researchers observed an improvement in all performance indices, with the median of the correlation
for the best model (neural networks) increasing from 0.75 to 0.82 and the RMSE dropping from 15µg/m3
to 7µg/m3 .
One of the main drawbacks of data-driven models for air quality is that they provide information
only in the point where the measurements are available. To overcome this limitation, recent literature has
presented mixed deterministic and data-driven approaches (see, for example, Carnevale et al., 2020) which
use the data assimilation procedure and offer promising forecasting performance.
From a practical point of view, critical issues regarding forecasting air quality include:

129 This subsection was written by Claudio Carnevale.

124
 • Information collection and data access: even if regional authorities have to publicly provide data and
information related to air quality and meteorology, the measured data are not usually available in
real-time and the interfaces are sometimes not automated;
• Data quantity: the amount of information required by air quality forecasting systems is usually large,
in particular towards the definition of the training and validation sets;
• Non-linear relationships: the phenomenon of accumulation of pollutants in atmosphere is usually
affected by strong nonlinearities, which significantly impact the selection of the models and their
performance;
• Unknown factors: it is a matter of fact that the dynamic of pollutants in atmosphere is affected by a
large number of non-measurable variables (such as meteorological variables or the interaction with
other non-measurable pollutants), largely affecting the capability of the models to reproduce the state
of the atmosphere.

3.5.4. Forecasting and decision making for floods and water resources management130
In Water Resources and Flood Risk Management, decision makers are frequently confronted with the
need of taking the most appropriate decisions not knowing what will occur in the future. To support
their decision-making under uncertainty, decision theory (Berger, 1985; Bernardo, 1994; DeGroot, 2004)
invokes Bayesian informed decision approaches, which find the most appropriate decision by maximis-
ing (or minimising) the expected value of a “utility function”, thus requiring its definition, together with
the estimation of a “predictive probability” density (Berger, 1985) due to the fact that utility functions
are rarely linear or continuous. Consequently, their expected value does not coincide with the value as-
sumed on the predicted “deterministic” expected value. Accordingly, overcoming the classical 18th century
“mechanistic” view by resorting into probabilistic forecasting approaches becomes essential.
The failure of decision-making based on deterministic forecasts in the case of Flood Risk Management
is easily shown through a simple example. At a river section, the future water level provided by a forecast
is uncertain and can be described by a Normal distribution with mean 10 meters and standard deviation
of 5 meters. Given a dike elevation of 10.5 meters, damages may be expected as zero if water level falls
below the dike elevation and linearly growing when level exceeds it with a factor of 106 dollars. If one
assumes the expect value of forecast as the deterministic prediction to compute the damage the latter will
result equal to zero, while if one correctly integrates the damage function times the predictive density the
estimated expected damage will results into 6.59 millions of dollars and educated decisions on alerting or
not the population or evacuating or not a flood-prone area can be appropriately taken.
Water resources management, and in particular reservoirs management, aim at deriving appropriate
operating rules via long term expected benefits maximisation. Nonetheless, during flood events decision
makers must decide how much to preventively release from multi-purpose reservoirs in order to reduce
dam failure and downstream flooding risks the optimal choice descending from trading-off between loos-
ing future water resource vs the reduction of short term expected losses.
This is obtained by setting up an objective function based on the linear combination of long and short
term “expected losses”, once again based on the available probabilistic forecast. This Bayesian adaptive
reservoir management approach incorporating into the decision mechanism the forecasting information

130 This subsection was written by Ezio Todini.

125
described by the short-term predictive probability density, was implemented on the lake Como since 1997
(Todini, 1999, 2017) as an extension of an earlier original idea (Todini, 1991). This resulted into:

• a reduction of over 30% of of the city of Como frequency;

• an average reduction of 12% of the water deficit;
• an increase of 3% in the electricity production.

Lake Como example clearly shows that instead of basing decisions on the deterministic prediction,
the use of a Bayesian decision scheme, in which model forecasts describe the predictive probability den-
sity, increases the reliability of the management scheme by essentially reducing the probability of wrong
decisions (Todini, 2017, 2018).

3.6. Social good and demographic forecasting

3.6.1. Healthcare131
There are many decisions that depend on the quality of forecasts in the health care system, from capac-
ity planning to layout decisions to the daily schedules. In general, the role of forecasting in health care is
to inform both clinical and non-clinical decisions. While the former concerns decisions related to patients
and their treatments (Makridakis et al., 2019), the latter involves policy/management, and supply chain
decisions that support the delivery of high-quality care for patients.
A number of studies refer to the use of forecasting methods to inform clinical decision making. These
methods are used to screen high risk patients for preventative health care (Chen et al., 2015; van der Mark
et al., 2014; Santos et al., 2015; Uematsu et al., 2014), to predict mental health issues (Shen et al., 2017;
Tran et al., 2013), to assist diagnosis and disease progression (Ghassemi et al., 2015; Ma et al., 2017; Pierce
et al., 2010; Qiao et al., 2019), to determine prognosis (Dietzel et al., 2010; Ng et al., 2007), and to recom-
mend treatments for patients (Kedia and Williams, 2003; Scerri et al., 2006; Shang et al., 2019). Common
forecasting methods to inform clinical decisions include time series, regression, classification tree, neu-
ral networks, Markov models and Bayesian networks. These models utilise structured and unstructured
data including clinician notes (Austin and Kusumoto, 2016; Labarere et al., 2014) which makes the data
pre-processing a crucial part of the forecasting process in clinical health care.
One of the aspects of the non-clinical forecasting that has received the most attention in both research
and application is the policy and management. Demand forecasting is regularly used in Emergency Depart-
ments (Arora et al., 2020; Choudhury and Urena, 2020; Khaldi et al., 2019; Rostami-Tabar and Ziel, 2020),
ambulance services (Al-Azzani et al., 2020; Setzler et al., 2009; Vile et al., 2012; Zhou and Matteson, 2016)
and hospitals with several different specialities (McCoy et al., 2018; Ordu et al., 2019; Zhou et al., 2018) to
inform operational, tactical and strategic planning. The common methods used for this purpose include
classical ARIMA and exponential smoothing methods, regression, singular spectrum analysis, Prophet,
Double-Seasonal Holt-Winter, TBATS and Neural Networks. In public health, forecasting can guide policy
and planning. Although it has a wider definition, the most attention is given to Epidemic forecasting (see
also section 3.6.2).
Forecasting is also used in both national and global health care supply chains, not only to ensure the
availability of medical products for the population but also to avoid excessive inventory. Additionally, the

131 This section was written by Bahman Rostami-Tabar.

126
lack of accurate demand forecast in a health supply chain may cost lives (Baicker et al., 2012) and has
exacerbated risks for suppliers (Levine et al., 2008). Classical exponential smoothing, ARIMA, regression
and Neural Network models have been applied to estimate the drug utilisation and expenditures (Dolgin,
2010; Linnér et al., 2020), blood demand (Fortsch and Khapalova, 2016), hospital supplies (Gebicki et al.,
2014; Riahi et al., 2013) and demand for global medical items (Amarasinghe et al., 2010; Hecht and Gandhi,
2008; van der Laan et al., 2016). It is important to note that, while the demand in a health care supply chain
has often grouped and hierarchical structures (Mircetica et al., 2020), this has not been well investigated
and needs more attention.

3.6.2. Epidemics and pandemics132

Forecasting the evolution of a pandemic, the growth of cases and fatalities for various horizons and
levels of granularity, is a complex task with raw and limited data – as each pandemic has unique features
with several factors affecting the severity and the contagiousness. Be that as it may, forecasting becomes
an paramount task for the countries to prepare and plan their response (Nikolopoulos, 2020), both in
healthcare and the supply chains (Beliën and Forcé, 2012).
Successful forecasting methods for the task include time-series methods, epidemiological and agent-
based models, metapopulation models, approaches in metrology (Nsoesie et al., 2013), machine and deep
learning methods (Yang et al., 2020). Andersson et al. (2008) used regression models for the prediction
of the peak time and volume of cases for a pandemic with evidence from seven outbreaks in Sweden.
Yaffee et al. (2011) forecasted the evolution of the Hantavirus epidemic in USA and compared causal and
machine-learning methods with time-series methods and found that univariate methods quite successful.
Soebiyanto et al. (2010) used ARIMA models for successfully short-term forecasting of influenza weekly
cases. Shaman and Karspeck (2012) used Kalman filter based SIR epidemiological models to forecast the
peak time of influenza 6-7 weeks ahead.
For COVID-19, Petropoulos and Makridakis (2020) applied a multiplicative exponential smoothing
model for predicting global number of confirmed cases, with very successful results both for point forecasts
and prediction intervals. This article got serious traction with 90,000 views and 140 citations in the first
six months since its publication, thus evidencing the importance of such empirical investigations. There
has been a series of studies focusing on predicting deaths in the USA and European countries for the first
wave of the COVID-19 pandemic (IHME COVID-19 health service utilization forecasting team and Murray,
2020a,b). Furthermore, Petropoulos et al. (2020) expanded their investigation to capture the continuation
of both cases and deaths as well as their uncertainty, achieving high levels of forecasting accuracy for ten-
days-ahead forecasts over a period of four months. Along the same lines, Doornik et al. (2020b) have been
publishing real-time accurate forecasts of confirmed cases and deaths from mid-March 2020 onwards.
Their approach is based on extraction of trends from the data using machine learning.
Pinson and Makridakis (2020) organised a debate between Taleb and Ioannidis on forecasting pan-
demics. Ioannidis et al. (2020) claim that forecasting for COVID-19 has by and large failed. However they
give recommendations of how this can be averted. They suggest that the focus should be on predictive
distributions and models should be continuously evaluated. Moreover, they emphasise the importance of
multiple dimensions of the problem (and its impact). Taleb et al. (2020) discuss the dangers of using naive,

132 This subsection was written by Konstantinos Nikolopoulos.

127
empirical approaches for fat-tailed variables and tail risk management. They also reiterate the inefficiency
of point forecasts for such phenomena.
Finally, Nikolopoulos et al. (2020) focused on forecast-driven planning, predicting the growth of COVID-
19 cases and the respective disruptions across the supply chain at country level with data from the USA,
India, UK, Germany, and Singapore. Their findings confirmed the excess demand for groceries and elec-
tronics, and reduced demand for automotive – but the model also proved that the earlier a lock-down is
imposed, the higher the excess demand will be for groceries. Therefore, governments would need to se-
cure high volumes of key products before imposing lock-downs; and, when this is not possible, seriously
consider more radical interventions such as rationing.

3.6.3. Forecasting dengue incidence133

Dengue is a mosquito-borne viral infection. Estimates of World Health Organisation reveals that that
about half of the world’s population is now at risk for Dengue infection. Aedes aegypti and Aedes albopictus
are the principal vector for dengue transmission and is a highly domesticated mosquito. The tropical and
sub-tropical areas form an ideal condition for them to be active all year around. Rainfall, temperature,
and relative humidity are thought as important factors attributing towards the growth and dispersion of
mosquito vectors and potential of dengue outbreaks. Forecasting dengue incidence can help public health
officials in many ways: (i) guide effective planning of medical services, (ii) improve prevention and control
programs, and (iii) analyse the effectiveness of strategies are to name few.
In reviewing the existing literature, two data types have been used to forecast dengue incidence: (i) in-
cidence of laboratory-confirmed dengue cases among the clinically suspected patients in various time fre-
quencies (such as weekly, monthly, quarterly, or yearly), and (ii) web-based data (Google trends, social
media) associated with Dengue cases. Dengue incidence data show complex nonlinear dynamics with
strong seasonality, nonstationarity (changes in dominant periodic components over time) and, sometimes,
intermittent dynamics.
SARIMA models have been quite popular in forecasting laboratory-confirmed dengue cases (Gharbi
et al., 2011; Martinez and Silva, 2011; Promprou et al., 2006). Chakraborty et al. (2019) used a hybrid
model combining ARIMA and neural network autoregressive (NNAR) to forecast dengue cases. In light
of the biological relationships between climate and transmission potential, several studies have used addi-
tional covariates such as, rainfall, temperature, wind speed, and humidity to forecasts dengue incidence.
Poisson regression model has been widely used to forecast dengue incidence using climatic factors and
lagged time between dengue incidence and weather variables (Hii et al., 2012; Koh et al., 2018). Several
researchers looked at the use of Quasi-Poisson and negative binomial regression models to accommodate
over dispersion in the counts (Lowe et al., 2011; Wang et al., 2014). Cazelles et al. (2005) used wavelet
analysis to explore the dynamic of dengue incidence and wavelet coherence analyses was used to identify
time and frequency specific association with climatic variables. de Almeida Marques-Toledo et al. (2017)
took a different perspective and look at weekly tweets to forecast Dengue cases. Rangarajan et al. (2019)
used Google trend data to forecast Dengue cases. Authors hypothesised that web query search related to
dengue disease correlated with the current level of dengue cases and thus may be helpful in forecasting
dengue cases.

133 This subsection was written by Thiyanga S. Talagala.

128
 A future direction to research in this field is to explore the use of spatio-temporal hierarchical forecast-
ing (Hyndman et al., 2011; Kourentzes and Athanasopoulos, 2019); see also section 2.9 on forecasting by
aggregation.

3.6.4. Forecasting mortality data134

Actuarial, Demographic and Health studies are some examples where mortality data are commonly
used. A valuable source of mortality information is the Human Mortality Database (HMD), a database that
provides mortality and population data for several countries or regions. In some situations, the lack of
reliable mortality data can be a problem, especially in developing countries, due to delays in registering
or miscounting deaths (Checchi and Roberts, 2005). Analysis of National Causes of Death for Action
(ANACONDA) is a valuable tool that assesses the accuracy and completeness of data for mortality and
cause of death by checking for potential errors and inconsistencies (Mikkelsen et al., 2020). The analysis
of mortality data is fundamental to public health authorities and policymakers, to make decisions or to
evaluate the effectiveness of prevention and response strategies adopted.
When facing a new pandemic, mortality surveillance is essential for monitoring the overall impact on
public health in terms of disease severity and mortality (Setel et al., 2020; Vestergaard et al., 2020). A
useful metric is excess mortality and is the difference between the observed number of deaths and the
expected number of deaths under “normal” conditions (Checchi and Roberts, 2005; Aron and Muellbauer,
2020). Thus, it can only be estimated with accurate and high-quality data from previous years. Excess
mortality has been used to measure the impact of heat events (Matte et al., 2016; Limaye et al., 2018),
pandemic influenza (Nunes et al., 2011; Nielsen et al., 2013) and nowadays COVID-19 (Nogueira et al.,
2020; Sinnathamby et al., 2020; Ritchie et al., 2020), among others. Excess mortality data have been making
available by the media publications The Economist, The New York Times and The Financial Times. Moreover,
a monitoring system of the weekly excess mortality in Europe has been performed by the EuroMOMO
project (Vestergaard et al., 2020).
Over the years, several methodologies were developed for mortality modelling and forecasting (Booth
and Tickle, 2008; Janssen, 2018). Amongst the three forecasting approaches (expectation, extrapolation,
and explanation), the majority of the developments were in the field of extrapolative methods. These
are considered more objective, easy to use and more likely to obtain better forecast accuracy than the
other two approaches (Janssen, 2018). The Lee-Carter mortality methodology (Lee and Carter, 1992) and
its variants have been one of the most widely used extrapolative methods. Lee-Carter methodology has
several advantages: it comprises a simple stochastic model with only one time-varying parameter, it per-
forms relatively well when past trends are linear, and it can forecast a changing age pattern of mortality
(Booth and Tickle, 2008). Extrapolative time series models are also commonly used, such as univariate
and multivariate ARIMA modelling. The introduction of stochastic time series models has the advantage
of obtaining a forecast probability distribution rather than a deterministic point forecast and, also, enable
the determination of forecast intervals (Booth and Tickle, 2008). Janssen (2018) presents a review of the
advances in mortality forecasting and possible challenges for future research.

134 This subsection was written by Clara Cordeiro.

129
3.6.5. Annuity price forecasting135
An important use of mortality forecasts for those individuals at age over 60 is in the pension and
insurance industries, whose profitability and solvency crucially rely on accurate mortality forecasts to
adequately hedge longevity risks (see, e.g., Shang and Haberman, 2020a,b). Longevity risk is a potential
systematic risk attached to the increasing life expectancy of annuitants, which can eventually result in
a higher payout ratio than expected (Crawford et al., 2008). When a person retires, an optimal way of
guaranteeing one individual’s financial income in retirement is to purchase an annuity (Yaari, 1965).
An annuity is a financial contract offered by insurers guaranteeing a steady stream of payments for
either a temporary or the lifetime of the annuitants in exchange for an initial premium fee. Temporary
annuities have grown in popularity in many countries, e.g., Australia and the United States of America.
Immediate lifetime annuities, where rates are locked in for life, have been shown to deliver poor value for
money (i.e., they may be expensive for the annuitants; see, for example, Cannon and Tonks, 2008, Chapter
6). These temporary annuities pay a pre-determined and guaranteed level of income which is higher than
the level of income provided by a lifetime annuity for a similar premium. Temporary annuities also offer
an alternative to lifetime annuities and allow the annuitants the option of also buying a deferred annuity
at a later date.
The price of an annuity with a maturity period is a random variable, as it depends on the value of zero-
coupon bond price and mortality forecasts. The zero-coupon bond price is a function of interest rate and
is comparably more stable than the mortality forecasts for the retirees.
In actuarial science and demography, mortality forecasting methods can be grouped into three cate-
gories: extrapolation, explanation and expectation (Booth and Tickle, 2008). The extrapolation approach
identifies age patterns and trends in time. The explanation approach captures the correlation between
mortality and underlying cause of death. The expectation approach is based on the subjective opinion of
experts, who set a long-run mortality target.
Methods based on expectation make use of the opinions of experts concerning future mortality or life
expectancy with a specified path and speed of progression towards the assumed value (Continuous Mor-
tality Investigation, 2020). The advantage of this approach is that demographic, epidemiological, medical
and other relevant information may be incorporated into the forecasts. The disadvantages are that such
information is subjective and biased towards experts’ opinions, and it only produces scenario-based deter-
ministic forecasts (Ahlburg and Vaupel, 1990; Wong-Fupuy and Haberman, 2004).
Methods based on explanation approach incorporate medical, social, environmental and behavioural
factors into mortality modelling. Example include smoking and disease-related mortality models. The
benefit of this approach is that mortality change can be understood from changes in related explanatory
variables; thus, it is attractive in terms of interpretability (Gutterman and Vanderhoof, 1998).
In the extrapolation approach, many parametric and nonparametric methods have been proposed (see,
e.g., Alho and Spencer, 2005; Hyndman and Ullah, 2007; Shang et al., 2011). Among the parametric meth-
ods, the method of Heligman and Pollard (1980) is well-known. Among the nonparametric methods, the
Lee-Carter model (Lee and Carter, 1992), Cairns-Blake-Dowd model (Cairns et al., 2009; Dowd et al.,
2010), and functional data model (Hyndman and Ullah, 2007), as well as their extensions and generalisa-
tions are dominant.

135 This subsection was written by Han Lin Shang.

130
3.6.6. Forecasting risk for violence and wars136
Can we predict the occurrence of WW3 in the next 20 years? Is there any trend in the severity of wars?
The study of armed conflicts and atrocities, both in terms of frequency over time and the number of ca-
sualties, has received quite some attention in the scientific literature and the media (e.g., Cederman, 2003;
Friedman, 2015; Hayes, 2002; Norton-Taylor, 2015; Richardson, 1948, 1960), falling within the broader
discussion about violence (Berlinski, 2009; Goldstein, 2011; Spagat et al., 2009), with the final goal of
understanding whether humanity is becoming less belligerent (Pinker, 2011), or not (Braumoeller, 2019).
Regarding wars and atrocities, the public debate has focused its attention on the so-called Long Peace
Theory (Gaddis, 1989), according to which, after WW2, humanity has experienced the most peaceful pe-
riod in history, with a decline in the number and in the severity of bloody events. Scholars like Mueller
(2009a,b) and Pinker (2011, 2018) claim that sociological arguments and all statistics suggest we live in
better times, while others like Gray (2015a,b) and Mann (2018) maintain that those statistics are often
partial and misused, the derived theories weak, and that war and violence are not declining but only being
transformed. For Mann, the Long Peace proves to be ad-hoc, as it only deals with Western Europe and
North America, neglecting the rest of the world, and the fact that countries like the US have been involved
in many conflicts out of their territories after WW2.
Recent statistical analyses confirm Gray’s and Mann’s views: empirical data do not support the idea
of a decline in human belligerence (no clear trend appears), and in its severity. Armed conflicts show
long inter-arrival times, therefore a relative peace of a few decades means nothing statistically (Cirillo and
Taleb, 2016b). Moreover, the distribution of war casualties is extremely fat-tailed (Clauset, 2018; Clauset
and Gleditsch, 2018), often with a tail exponent ξ > 1 (Cirillo and Taleb, 2016b), indicating a possibly
infinite mean, i.e., a tremendously erratic and unforeseeable phenomenon. An only apparently infinite-
mean phenomenon though (Cirillo and Taleb, 2019), because no single war can kill more than the entire
world population, therefore a finite upper bound exists, and all moments are necessarily finite, even if
difficult to estimate. Extreme value theory (Embrechts et al., 2013) can thus be used to correctly model tail
risk and make prudential forecasts (with many caveats like in Scharpf et al., 2014), while avoiding naive
extrapolations (Taleb et al., 2020).
As history teaches (Nye, 1990), humanity has already experienced periods of relative regional peace,
like the famous Paces Romana and Sinica. The present Pax Americana is not enough to claim that we are
structurally living in a more peaceful era. The Long Peace risks to be another apophenia, another example
of Texan sharpshooter fallacy (Carroll, 2003).
Similar mistakes have been made in the past. Buckle (1858) wrote: “that [war] is, in the progress
of society, steadily declining, must be evident, even to the most hasty reader of European history. If we
compare one country with another, we shall find that for a very long period wars have been becoming less
frequent; and now so clearly is the movement marked, that, until the late commencement of hostilities, we
had remained at peace for nearly forty years: a circumstance unparalleled [...] in the affairs of the world”.
Sadly, Buckle was victim of the illusion coming from the Pax Britannica (Johnston, 2008): the century
following his prose turned out to be the most murderous in human history.

136 This subsection was written by Pasquale Cirillo.

131
3.7. Systems and humans
3.7.1. Support systems137
Forecasting in businesses is a complicated procedure, especially when predicting numerous, diverse
series, dealing with unstructured data of multiple sources, and incorporating human judgment (Lim and
O’Connor, 1996a). In this respect, since the early 80’s, various Forecasting Support Systems (FSSs) have
been developed to facilitate forecasting and support decision making (Kusters et al., 2006). Rycroft (1993)
provides an early comparative review of such systems, while many studies strongly support their utilisa-
tion over other forecasting alternatives (Tashman and Leach, 1991; Sanders and Manrodt, 2003).
In a typical use-case scenario, the FSSs will retrieve the data required for producing the forecasts, will
provide some visualisations and summary statistics to the user, allow for data pre-processing, and then
produce forecasts that may be adjusted according to the preferences of the user. However, according to
Ord and Fildes (2013), effective FSS should be able to produce forecasts by combining relevant information,
analytical models, judgment, visualisations, and feedback. To that end, FSSs must (i) elaborate accurate,
efficient, and automatic statistical forecasting methods, (ii) enable users to effectively incorporate their
judgment, (iii) allow the users to track and interact with the whole forecasting procedure, and (iv) be easily
customised based on the context of the company.
Indeed, nowadays, most off-the-self solutions, such as SAP, SAS, JDEdwards, and ForecastPro, offer
a variety of both standard and advanced statistical forecasting methods, as well as data pre-processing
and performance evaluation algorithms. On the other hand, many of them still struggle to incorporate
state-of-the-art methods that can further improve forecasting accuracy, such as automatic model selection
algorithms and temporal aggregation (Petropoulos, 2015), thus limiting the options of the users. Similarly,
although many FSSs support judgmental forecasts and judgmental adjustments of statistical forecasts,
this is not done as suggested by the literature, i.e., in a guided way under a well-organised framework.
As a result, the capabilities of the users are restrained and methods that could be used to mitigate biases,
overshooting, anchoring, and unreasonable or insignificant changes that do not rationalise the time wasted,
are largely ignored (Fildes and Goodwin, 2013; Fildes et al., 2006).
Other practical issues of FSSs are related with their engine and interfaces which are typically de-
signed so that they are generic and capable to serve different companies and organisations of diverse needs
(Kusters et al., 2006). From a developing and economic perspective, this is a reasonable choice. However,
the lack of flexibility and customisability can lead to interfaces with needless options, models, tools, and
features that may confuse inexperienced users and undermine their performance (Fildes et al., 2006). Thus,
simple, yet exhaustive interfaces should be designed in the future to better serve the needs of each com-
pany and fit its particular requirements (Spiliotis et al., 2015). Ideally, the interfaces should be adapted
to the strengths and weaknesses of the user, providing useful feedback when possible (Goodwin et al.,
2007). Finally, web-based FSSs could replace windows-based ones that are locally installed and therefore
of limited accessibility, availability, and compatibility (Asimakopoulos and Dix, 2013). Cloud computing
and web-services could be exploited in that direction.

137 This subsection was written by Vassilios Assimakopoulos.

132
3.7.2. Cloud resource capacity forecasting138
One of the central promises in cloud computing is that of elasticity. Customers of cloud computing
services can add compute resources in real-time to meet and satisfy increasing demand and, when demand
for a cloud-hosted application goes down, it is possible for cloud computing customers to down-scale.
The benefit of the latter is particularly economically interesting during the current pandemic. Popular
recent cloud computing offerings take this elasticity concept one step further. They abstract away the
computational resources completely from developers, so that developers can build serverless applications.
In order for this to work, the cloud provider handles the addition and removal of compute resources
“behind the scenes”.
To keep the promise of elasticity, a cloud provider must address a number of forecasting problems at
varying scales along the operational, tactical and strategic problem dimensions (Januschowski and Kolassa,
2019). As an example for a strategic forecasting problems: where should data centres be placed? In what
region of a country and in what geographic region? As an example for tactical forecasting problems, these
must take into account energy prices (see Section 3.4.3) and also, classic supply chain problems (Larson
et al., 2001). After all, physical servers and data centres are what enables the cloud and these must be
ordered and have a lead-time. The careful incorporation of life cycles of compute types is important (e.g.,
both the popularity of certain machine types and the duration of a hard disk). Analogous to the retail
sector, cloud resource providers have tactical cold-start forecasting problems. For example, while GPU or
TPU instances are still relatively recent but already well estabilished, the demand for quantum computing
is still to be decided. In the class of operational forecasting problems, cloud provider can choose to address
short-term resource forecasting problems for applications such as adding resources to applications predic-
tively and make this available to customers (Barr, 2018). The forecasting of the customer’s spend for cloud
computing is another example. For serverless infrastructure, a number of servers is often maintained in a
ready state (Gias and Casale, 2020) and the forecasting of the size of this ‘warmpool’ is another example.
We note that cloud computing customers have forecasting problems that mirror the forecasting challenges
of the cloud providers. Interestingly, forecasting itself has become a software service that cloud computing
companies offer (Januschowski et al., 2018a; Poccia, 2019; Liberty et al., 2020)
Many challenges in this application area are not unique to cloud computing. Cold start problems exist
elsewhere for example. What potentially stands out in cloud computing forecasting problems may be the
scale (e.g., there are a lot of physical servers available), the demands on the response time and granularity
of a forecast and the degree of automation. Consider the operational forecasting problem of predictive
scaling. Unlike in retail demand forecasting, no human operator will be able to control this and response
times to forecasts are in seconds. It will be interesting to see whether approaches based on reinforcement
learning (Gamble and Gao, 2018) can partially replace the need to have forecasting models (Januschowski
et al., 2018b).

3.7.3. Including judgment in forecasting practice139

Surveys of forecasting practice (De Baets, 2019) have shown that of use of pure judgmental forecasting
by practitioners has become less common. About 40 years ago, Sparkes and McHugh (1984) found that

138 This subsection was written by Tim Januschowski.

139 This subsection was written by Nigel Harvey.

133
company action was more likely to be influenced by judgment forecasts than by any other type of forecast.
In contrast, Fildes and Petropoulos (2015) found that nowadays only 15.6% of forecasts in the companies
that they surveyed were made by judgment alone. The majority of forecasts (55.6%) were made using a
combination of statistical and judgmental methods.
This begs the question of whether judgmental forecasting is still relevant and how wise it is to rely
on this type of forecasting? Answers here depend on the type of information on which the judgmental
forecasts are based (Harvey, 2007). Recently, a major study on geo-political forecasts based solely on infor-
mation held in the forecasters’ memory has been summarised by Tetlock and Gardner (2015). This showed
clear individual differences: a small proportion of participants (‘superforecasters’) were particularly good
at making probabilistic forecasts about geo-political matters. Furthermore, people could be trained to
make better probabilistic forecasts of this type.
In contrast, people have difficulty making cross-series forecasts. They have difficulty learning the cor-
relation between variables and using it to make their forecasts (Harvey et al., 1994; Lim and O’Connor,
1996a,b). Additionally, they appear to take account of the noise as well as the pattern when learning the
relation between variables; hence, when later using one of the variables to forecast the other, they add
noise to their forecasts (Gray et al., 1965).
Judgmental extrapolation from a single time series is subject to various effects. First, people are influ-
enced by optimism. For example, they over-forecast time series labelled as ‘profits’ but under-forecast the
same series labelled as ‘losses’ (Harvey and Reimers, 2013). Second, they add noise to their forecasts so
that a sequence of forecasts looks similar to (‘represents’) the data series (Harvey, 1995). Third, they damp
trends in the data (Eggleton, 1982; Harvey and Reimers, 2013; Lawrence and Makridakis, 1989). Fourth,
forecasts from un-trended independent series do not lie on the series mean but between the last data point
and the mean; this is what we would expect if people perceived a positive autocorrelation in the series
(Reimers and Harvey, 2011). These last two effects can be explained in terms of the under-adjustment that
characterises use of anchor-and-adjust heuristics: forecasters anchor on the last data point and adjust to-
wards the trend line or mean – but do so insufficiently. However, in practice, this under-adjustment may be
appropriate because real linear trends do become damped and real series are more likely to contain a mod-
est autocorrelation than be independent (Harvey, 2011). We should therefore be reluctant to characterise
these last two effects as biases.
In summary, the literature suggests that there are good reasons for practitioners to prefer statistical or
combined statistical and judgmental approaches to pure judgmental forecasting when past values of the
variable to be forecast (or past values of variables correlated with the variable to be forecast) are available.
When no such information is available, pure judgmental forecasting may provide the only option. In this
situation, selection and training of forecasters would be sensible.

3.7.4. Judgmental adjustments in forecasting practice140

Judgmental adjustments give practitioners a quick and convenient way to incorporate their insights,
experience and additional information they possess into a set of baseline forecasts that have been externally
produced. Their effectiveness has made them a highly cherished tool among forecasting practitioners
(Arvan et al., 2019; Eksoz et al., 2019; Lawrence et al., 2006; Petropoulos et al., 2016).

140 This subsection was written by M. Sinan Gönül.

134
 Recent research based on business practice and consultancy work have revealed intriguing findings
about the way that practitioners make judgmental adjustments. Fildes et al. (2009) examined the judg-
mental adjustment applications in four large supply-chain companies and found evidence that the adjust-
ments in a ‘negative’ direction improved the accuracy more than the adjustments in a ‘positive’ direction.
Analogous findings were observed by Franses and Legerstee (2009b) who examined a large set of forecasts
produced by a multinational pharmaceutical company. This effect may be attributable to wishful thinking
or optimism that may underlie positive adjustments.
The adjustments ‘larger’ in magnitude were also more beneficial in terms of the final forecast accuracy
than ‘smaller’ adjustments (Fildes et al., 2009). This may simply be because smaller adjustments are merely
a sign of tweaking the numbers, but large adjustments are conducted when there is a highly valid reason
to make them. Syntetos et al. (2009) also found supporting evidence for these tendencies in the case of
forecasting products with intermittent demand. The positive adjustments were less accurate than the
negative ones and smaller adjustments ended up producing worse performance than larger ones.
Given the undisputable existence of judgmental adjustments in forecasting practice what are the rea-
sons that practitioners report for introducing their adjustments? Önkal and Gönül (2005) conducted a
series of interviews and a survey on forecasting practitioners (Gönül et al., 2009) to explore these rea-
sons. The main reasons conveyed are (i) to incorporate the practitioners’ intuition and experience about
the predictions generated externally, (ii) to accommodate sporadic events and exceptional occasions, (iii)
to integrate confidential/insider information that may have not been captured in the forecasts, (iv) to hold
responsibility and to gain control of the forecasting process, (v) to incorporate the expectations and view-
points of the practitioners, and (vi) to compensate for various judgmental biases that are believed to exist
in the predictions. These studies also revealed that forecasting practitioners are very fond of judgmental
adjustments and perceive them as a prominent way of ‘completing’ and ‘owning’ the predictions that are
generated by others.
Conversely, what are the primary reasons for practitioners preferring not to adjust? Önkal and Gönül
(2005) and Gönül et al. (2009) report that these occasions are (i) when the practitioners are adequately in-
formed and knowledgeable about the forecasting method(s) that are used to generate the baseline forecasts,
(ii) when there are accompanying explanations and convincing communications that provide the rationale
behind forecast method selection, (iii) when baseline predictions are supplemented by additional support-
ive materials such as scenarios and alternative forecasts, (iv) when the forecasting source is believed to be
trustworthy and reliable, and (v) when organisational policy or culture prohibits judgmental adjustments.
In these circumstances, the baseline forecasts are accepted more by the practitioners and their adjustments
tend to be less frequent.
Despite the plethora of recent research on judgmental adjustments, there are still many mechanisms
to understand and many avenues to explore with respect to this popular intervention tool that serves
forecasting practitioners.

3.7.5. Supporting judgment in practice141

Forecasting practice relies on Forecast Support Systems (FSS) that combine judgment and statistics,
attempting to facilitate improved forecasting accuracy (Fildes et al., 2006). But how can we design FSS

141 This subsection was written by Shari De Baets.

135
for the successful integration of judgment and statistics? The perceived quality and accessibility of a FSS
can be influenced by its design. More on this can be found in the literature on the Technology Acceptance
Model (Davis et al., 1989) and decision making (for instance by means of framing, visual presentation or
nudging; e.g., Gigerenzer, 1996; Kahneman and Tversky, 1996; Payne, 1982; Thaler and Sunstein, 2009).
A number of studies have investigated the design aspects of FSS, with varying success. One of the more
straightforward approaches is to change the look and feel of the FSS as well as its presentation style. Harvey
and Bolger (1996) found that trends for instance were more easily discernible when the data was displayed
graphically rather than tabular. Additionally, simple variations in presentation such as line graphs versus
point graphs can alter accuracy (Theocharis et al., 2018).
One could also tweak the functionalities of the FSS. Goodwin (2000b) investigated three ways of im-
proving judgmental adjustment via changes in the FSS. The first change was to set ‘no adjusting’ as default
in the program. This would make ‘no change’ the default and ‘change’ the more effortful option. The
second change is programming the FSS so that it would ask for the adjustment size, if the forecaster indi-
cated a willingness to adjust. A third implementation was the addition of an ‘explanation’ feature: when
the forecaster wanted to adjust the forecast, they were explicitly asked to document the reason for the
change. The default option and the explanation feature were successful in heightening the acceptance of
the statistical forecast and thus improving the forecast accuracy. Asking for the size of the adjustment was
not. Other ways of influencing forecaster’s behaviour with the FSS is to provide guidance or restriction
(Goodwin et al., 2011). They set up an experiment such that one group received ‘guidance’ in the form
of information, and the other group was ‘restricted’ in what they could do. The latter group could, for
instance, not make small adjustments, which have previously been found to be more damaging than they
are helpful (Fildes et al., 2009). However, the researchers found that neither restrictiveness nor guidance
was successful in improving accuracy, and both were met with resistance by the forecasters.
While the previous works focused on voluntary integration, Goodwin (2000a) also published a study on
the mechanical integration. In this case, the FSS processes the changes. This could be done via psycholog-
ical bootstrapping, combining different inputs automatically with equal or varying weights, or automatic
correction for bias by the FSS (Goodwin, 2000a, 2002). A way of automatically combining judgment and
statistics was investigated by Baecke et al. (2017), who compared normal judgmental adjustment with what
they term “integrative judgment”. In brief, integrative judgment takes the judgmental information into ac-
count as a predictive variable in the forecasting model and generates a new forecast. The advantage of this
method is that forecasters still have their input into the forecasting process and the resistance found in the
study by Goodwin et al. (2011) should not thus occur.
Whatever the solution will be, the FSS will have to improve forecast accuracy, but in addition, it will
need to be easy to use, understandable, and acceptable (Fildes et al., 2006).

3.7.6. Scoring expert forecasts for decision making142

Evaluating forecasting capabilities can be a difficult task. One prominent way to evaluate an expert’s
forecast is to score the forecast once the realisation of the uncertainty is known. Scoring forecasts using
the outcome’s realisations over multiple forecasts offers insights into an individual’s expertise. Experts can
also use scoring information to identify ways to improve future forecasts. In addition, scoring rules and

142 This subsection was written by Yael Grushka-Cockayne.

136
evaluation measures can be designed to match decision-making problems, incentivising forecasts that are
most useful in a specific situation (Winkler et al., 2019).
Scoring rules were first suggested for evaluate meteorological forecasts in work by Brier (1950). Scoring
rules have since been used in a wide variety of settings, such as business and other applications. When
forecasting a discrete uncertainty with only two possible outcomes (e.g., a loan with be defaulted on or
not, a customer will click on an ad or not), the Brier score assigns a score of −(1 − p)2 , where p is the
probability forecast reported that the event will occurs. The greater the probability reported for an event
that occurs, the higher the score the forecast receives. Over multiple forecasts, better forecasters will tend
to have higher average Brier scores. For discrete events with more than two outcomes, a logarithmic scoring
rule can be used.
The scoring rules are attractive to managers in practice since they are considered proper. Proper scoring
rules incentivise honest forecasts from the experts, even prior to knowing the realisation of an uncertainty,
since ex ante the expected score is maximised only when reported probabilities equals true beliefs (Winkler
et al., 1996; O’Hagan et al., 2006; Bickel, 2007; Gneiting and Raftery, 2007; Merkle and Steyvers, 2013).
Examples of a scoring rule that is not proper yet still commonly used are the linear score, which simply
equals the reported probability or density for the actual outcome, or the skill score, which is the percentage
improvement of the Brier score for the forecast relative to the Brier score of some base line naive forecast
(Winkler et al., 2019).
For forecasting continuous quantities, forecasts could be elicited by asking for an expert’s quantile
(or fractile) forecast rather than a probability forecast. For instance, the 0.05, 0.25, 0.50, 0.75 and 0.95
quantiles are often elicited in practice, and in some cases every 0.01 quantile, between 0-1 are elicited (e.g.,
the 2014 Global Energy Forecasting Competition, Hong et al., 2016). Proper scoring rules for quantiles are
developed in Jose and Winkler (2009).
When forecasts are used for decision-making, it is beneficial if the scoring rule used relates in some
manner to the decision problem itself. In certain settings, the connection of the scoring rule to the deci-
sion context is straight forward. For example, Jose et al. (2008) develop scoring rules that can be mapped
to decision problems based on the decision maker’s utility function. Johnstone et al. (2011) develop tai-
lored scoring rules aligning the interest of the forecaster and the decision maker. Grushka-Cockayne et al.
(2017b) link quantile scoring rules to business profit-sharing situations.
The use of prediction markets as a mechanism for collecting forecasting have been proposed as practical
and useful for aligning forecasting activities with business goals. Chen et al. (2005) found that prediction
markets can generate highly accurate prediction. Bassamboo et al. (2015) demonstrate the use of predic-
tion markets in forecasting quantities important to operational decisions, such as sales forecasts, price
commodity forecasts, or product features.

3.7.7. Trust in forecasts143

Regardless of how much effort is poured into training forecasters and developing elaborate forecast
support systems, decision-makers will either modify or discard the predictions if they do not trust them.
Hence, trust is essential for forecasts to be actually used in making decisions (Alvarado-Valencia and Bar-
rero, 2014; Önkal et al., 2019).

143 This subsection was written by Dilek Önkal.

137
 Given that trust appears to be the most important attribute that promotes a forecast, what does it mean
to practitioners? Past work suggests that trusting a forecast is often equated with trusting the forecaster,
their expertise and skills so that predictions could be used without adjustment to make decisions (Önkal
et al., 2019). It is argued that trust entails relying on credible forecasters that make the best use of available
information while using correctly applied methods and realistic assumptions (Gönül et al., 2009) with no
hidden agendas (Gönül et al., 2012). Research suggests that trust is not only about trusting forecaster’s
competence; users also need to be convinced that no manipulations are made for personal gains and/or to
mislead decisions (Twyman et al., 2008).
Surveys with practitioners show that key determinants of trust revolve around (i) forecast support fea-
tures and tools (e.g., graphical illustrations, rationale for forecasts), (ii) forecaster competence/credibility,
(iii) forecast combinations (from multiple forecasters/methods), and (iv) forecast user’s knowledge of fore-
casting methods (Önkal et al., 2019).
What can be done to enhance trust? If trust translates into accepting guidance for the future while
acknowledging and tolerating potential forecast errors, then both the providers and users of forecasts
need to work as partners towards shared goals and expectations. Important pathways to accomplish this
include (i) honest communication of forecaster’s track record and relevant accuracy targets (Önkal et al.,
2019), (ii) knowledge sharing (Özer et al., 2011; Renzl, 2008) and transparency of forecasting methods,
assumptions and data (Önkal et al., 2019), (iii) communicating forecasts in the correct tone and jargon-
free language to appeal to the user audience (Taylor and Thomas, 1982), (iv) users to be supported with
forecasting training (Merrick et al., 2006), (v) providing explanations/rationale behind forecasts (Gönül
et al., 2006; Önkal et al., 2008), (vi) presenting alternative forecasts under different scenarios (Önkal et al.,
2013), and (vii) giving combined forecasts as benchmarks (Önkal et al., 2019).
Trust must be earned and deserved (Maister et al., 2012) and is based on building a relationship that
benefits both the providers and users of forecasts. Take-aways for those who make forecasts and those who
use them converge around clarity of communication as well as perceptions of competence and integrity.
Key challenges for forecasters are to successfully engage with users throughout the forecasting process
(rather than relying on a forecast statement at the end) and to convince them of their objectivity and
expertise. In parallel, forecast users face challenges in openly communicating their expectations from
forecasts (Gönül et al., 2009), as well as their needs for explanations and other informational addendum to
gauge the uncertainties surrounding the forecasts. Organisational challenges include investing in forecast
management and designing resilient systems for collaborative forecasting.

3.7.8. Communicating forecast uncertainty144

Communicating forecast uncertainty is a critical issue in forecasting practice. Effective communica-
tion allows forecasters to influence end-users to respond appropriately to forecasted uncertainties. Some
frameworks for effective communication have been proposed by decomposing the communication process
into its elements: the communicator, object of uncertainty, format of expression, audience, and its effect
(National Research Council, 2006; van der Bles et al., 2019).
Forecasters have long studied part of this problem focusing mostly in the manner by which we express
forecast uncertainties. Gneiting and Katzfuss (2014) provides a review of recent probabilistic forecasting

144 This subsection was written by Victor Richmond R. Jose.

138
methods. Forecasting practice however revealed that numeracy skills and cognitive load can often inhibit
end-users from correctly interpreting these uncertainties (Joslyn and Nichols, 2009; Raftery, 2016). At-
tempts to improve understanding through the use of less technical vocabulary also creates new challenges.
Research in psychology show that wording and verbal representation play important roles in disseminating
uncertainty (Joslyn et al., 2009). Generally forecasters are found to be consistent in their use of terminol-
ogy, but forecast end-users often have inconsistent interpretation of these terms even those commonly
used (Budescu and Wallsten, 1985; Clark, 1990; Ülkümen et al., 2016). Pretesting verbal expressions and
avoiding commonly misinterpreted terms are some easy ways to significantly reduce biases and improve
comprehension.
Visualisations can also be powerful in communicating uncertainty. Johnson and Slovic (1995) and
Spiegelhalter et al. (2011) propose several suggestions for effective communication (e.g., multiple-format
use, avoiding framing bias, and acknowledging limitations), but also recognise the limited amount of
existing empirical evidence. Some domain-specific studies do exist. For example, Riveiro et al. (2014)
showed uncertainty visualisation helped forecast comprehension in a homeland security context.
With respect to the forecaster and her audience, issues such as competence, trust, respect, and optimism
have been recently examined as a means to improve uncertainty communication. Fiske and Dupree (2014)
discusses how forecast recipients often infer apparent intent and competence from the uncertainty pro-
vided and use these to judge trust and respect. This suggests that the amount of uncertainty information
provided should be audience dependent (Politi et al., 2007; Han et al., 2009). Raftery (2016) acknowledges
this by using strategies depending on the audience type (e.g., low-stakes user, risk avoider, etc.). Fischhoff
and Davis (2014) suggests a similar approach by examining how people are likely to use the information
(e.g., finding a signal, generating new options, etc.)
When dealing with the public, experts assert that communicating uncertainty helps users understand
forecasts better and avoid a false sense of certainty (Morss et al., 2008). Research however shows that
hesitation to include forecast uncertainty exists among experts because it provides an opening for criticism
and the possibility of misinterpration by the public (Fischhoff, 2012). This is more challenging when the
public has prior beliefs on a topic or trust has not been established. Uncertainty can be used by individuals
to reinforce a motivated-reasoning bias that allows them to “see what they want to see” (Dieckmann et al.,
2017). Recent work however suggests that increasing transparency for uncertainty does not necessarily
affect trust in some settings. van der Bles et al. (2020) recently showed in a series of experiments that
people recognise greater uncertainty with more information but expressed only a small decrease in trust
in the report and trustworthiness of the source.

3.8. Other applications

3.8.1. Tourism demand forecasting145
As seen throughout 2020, (leisure) tourism demand is very sensitive to external shocks such as natural
and human-made disasters, making tourism products and services extremely perishable (Frechtling, 2001).
As the majority of business decisions in the tourism industry require reliable demand forecasts (Song et al.,
2008), improving their accuracy has continuously been on the agenda of tourism researchers and practi-

145 This subsection was written by Ulrich Gunter.

139
tioners alike. Depending on data availability, as well as on geographical aggregation level, tourism demand
is typically measured in terms of arrivals, bed-nights, visitors, exports receipts, import expenditures, etc.
Since there are no specific tourism demand forecast models, standard univariate and multivariate sta-
tistical models, including common aggregation and combination techniques, etc., have been used in quan-
titative tourism demand forecasting (see, for example, Song et al., 2019; Jiao and Chen, 2019, for recent
reviews). Machine learning and other artificial intelligence methods, as well as hybrids of statistical and
machine learning models, have recently been employed more frequently.
Traditionally, typical micro-economic demand drivers (own price, competitors’ prices, and income)
and some more tourism-specific demand drivers (source-market population, marketing expenditures, con-
sumer tastes, habit persistence, and dummy variables capturing one-off events or qualitative characteris-
tics) have been employed as predictors in tourism demand forecasting (Song et al., 2008). One caveat of
some of these economic demand drivers is their publication lag and their low frequency, for instance, when
real GDP (per capita) is employed as a proxy for travellers’ income.
The use of leading indicators, such as industrial production as a leading indicator for real GDP, has
been proposed for short-term tourism demand forecasting and nowcasting (Chatziantoniou et al., 2016).
During the past couple of years, web-based leading indicators have also been employed in tourism demand
forecasting and have, in general, shown improvement in terms of forecast accuracy. However, this has
not happened in each and every case, thereby confirming the traded wisdom that there is no single best
tourism demand forecasting approach (Li et al., 2005). Examples of those web-based leading indicators
include Google Trends indices (Bangwayo-Skeete and Skeete, 2015), Google Analytics indicators (Gunter
and Önder, 2016), as well as Facebook ‘likes’ (Gunter et al., 2019).
The reason why these expressions of interaction of users with the Internet have proven worthwhile as
predictors in a large number of cases is that it is sensible to assume potential travellers gather information
about their destination of interest prior to the actual trip, with the Internet being characterised by com-
parably low search costs, ergo allowing potential travellers to forage information (Pirolli and Card, 1999)
with only little effort (Zipf, 2016). A forecaster should include this information in their own set of relevant
information at the forecast origin (Lütkepohl, 2005), if taking it into account results in an improved fore-
cast accuracy, with web-based leading indicators thus effectively Granger-causing (Granger, 1969) actual
tourism demand.
Particularly during the first half of the COVID-19 pandemic, which was characterised by global travel
restrictions and tourism businesses being locked down, scenario forecasting and other forms of hybrid and
judgmental forecasting played an important role, thereby highlighting an important limitation of quanti-
tative tourism demand forecasting as currently practised. Based on the rapid development of information
technology and artificial intelligence, Li and Jiao (2020), however, envisage a “super-smart tourism fore-
casting system” (Li and Jiao, 2020, p. 264) for the upcoming 75 years of tourism demand forecasting.
According to these authors, this system will be able to automatically produce forecasts at the micro level
(i.e., for the individual traveller and tourism business) in real time while drawing on a multitude of data
sources and integrating multiple (self-developing) forecast models.

140
3.8.2. Forecasting for airports146
Airports have long invested in forecasting arrivals and departures of aircrafts. These forecasts are im-
portant in measuring airspace and airport congestions, designing flight schedules, and planning for the
assignment of stands and gates (Barnhart and Cohn, 2004). Various techniques have been applied to fore-
cast aircrafts’ arrivals and departures. For instance, Rebollo and Balakrishnan (2014) apply random forests
to predict air traffic delays of the National Airspace System using both temporal and network delay states
as covariates. Manna et al. (2017) develop a statistical model based on a gradient boosting decision tree
to predict arrival and departure delays, using the data taken from the United States Department of Trans-
portation (Bureau of Transportation Statistics, 2020). Rodrı́guez-Sanz et al. (2019) develop a Bayesian
Network model to predict flight arrivals and delays using the radar data, aircraft historical performance
and local environmental data. There are also a few studies that have focused on generating probabilistic
forecasts of arrivals and departures, moving beyond point estimates. For example, Tu et al. (2008) develop
a predictive system for estimating flight departure delay distributions using flight data from Denver In-
ternational Airport. The system employs the smoothing spline method to model seasonal trends and daily
propagation patterns. It also uses mixture distributions to estimate the residual errors for predicting the
entire distribution.
Passenger-centric forecasting problems have received some attention in the literature, particularly in
the past decade. Wei and Hansen (2006) build an aggregate demand model for air passenger traffic in a
hub-and-spoke network. The model is a log-linear regression that uses airline service variables such as
aircraft size and flight distance as predictors. Barnhart et al. (2014) develop a multinomial logit regression
model, designed to predict delays of US domestic passengers. Their study also uses data from the US De-
partment of Transportation (Bureau of Transportation Statistics, 2020). Guo et al. (2020) recently develop
a predictive system that generates distributional forecasts of connection times for transfer passengers at an
airport, as well as passenger flows at the immigration and security areas. Their approach is based on the
application of regression trees combined with copula-based simulations. This predictive system has been
implemented at Heathrow airport since 2017.
With an increasing amount of available data that is associated with activities in the aviation industry,
predictive analyses and forecasting methods face new challenges as well as opportunities, especially in
regard to updating forecasts in real time. The predictive system developed by Guo et al. (2020) is able
to generate accurate forecasts using real-time flight and passenger information on a rolling basis. The
parameters of their model, however, do not update over time. Therefore, a key challenge in this area is for
future studies to identify an efficient way to dynamically update model parameters in real time.

3.8.3. Traffic flow forecasting147

Traffic flow forecasting is an important task for traffic management bodies to reduce traffic congestion,
perform planning and allocation tasks, as well as for travelling individuals to plan their trips. Traffic flow
is complex spatial and time-series data exhibiting multiple seasonalities and affected by spatial exogenous
influences such as social and economic activities and events, various government regulations, planned road
works, weather, traffic accidents, etc. (Polson and Sokolov, 2017).

146 This subsection was written by Xiaojia Guo.

147 This subsection was written by Alexander Dokumentov.

141
 Methods to solve traffic flow forecasting problems vaguely fall into three categories. The first uses
parametric statistical methods such as ARIMA, seasonal ARIMA, space-time ARIMA, Kalman filters, etc.
(see, for example, Whittaker et al., 1997; Vlahogianni et al., 2004; Kamarianakis and Prastacos, 2005;
Vlahogianni et al., 2014). The second set of approaches uses purely of neural networks (Mena-Oreja and
Gozalvez, 2020). The third group of methods uses various machine learning, statistical non-parametric
techniques or mixture of them (see, for example, Hong, 2011; Zhang et al., 2016, 2017).
Although neural networks are probably the most promising technique for traffic flow forecasting (see,
for example, Polson and Sokolov, 2017; Do et al., 2019), statistical techniques, such as Seasonal-Trend
decomposition based on Regression (STR), can outperform when little data is available or they can be used
for imputation, de-noising, and other pre-processing before feeding data into neural networks which often
become less powerful when working with missing or very noisy data.
Traffic flow forecasting is illustrated below using vehicle flow rate data from road camera A1.GT.24538
on A1 highway in Luxembourg (La Fabrique des Mobilités, 2020) from 2019-11-19 06:44:00 UTC to 2019-
12-23 06:44:00 UTC. Most of the data points are separated by 5 minutes intervals. Discarding points
which do not follow this schedule leads to a data set where all data points are separated by 5 minutes
intervals, although values at some points are missing. The data is split into training and test sets by setting
aside last 7 days of data. As Hou et al. (2014) and Polson and Sokolov (2017) suggest, spatial factors are
less important for long term traffic flow forecasting, and therefore they are not taken into account and
only temporal data is used. Application of STR (Dokumentov, 2017) as a forecasting technique to the log
transformed data leads to a forecast with Mean Squared Error 102.4, Mean Absolute Error 62.8, and Mean
Absolute Percentage Error (MAPE) 14.3% over the test set, outperforming Double-Seasonal Holt-Winters
by 44% in terms of MAPE. The decomposition and the forecast obtained by STR are shown on figure 15
and the magnified forecast and the forecasting errors are on figure 16.

3.8.4. Call arrival forecasting148

Forecasting of inbound call arrivals for call centres supports a number of key decisions primarily
around staffing (Aksin et al., 2007). This typically involves matching staffing level requirements to ser-
vice demand as summarized in Figure 17. To achieve service level objectives, an understanding of the call
load is required in terms of the call arrivals (Gans et al., 2003). As such, forecasting of future call volume
or call arrival rates is an important part of call centre management.
There are several properties to call arrival data. Depending on the level of aggregation and the fre-
quency with which data is collected, e.g., hourly, call arrival data may exhibit intraday (within-day), in-
traweek, and intrayear multiple seasonal patterns (Avramidis et al., 2004; Brown et al., 2005b). In addition,
arrival data may also exhibit interday and intraday dependencies, with different time periods within the
same day, or across days within the same week, showing strong levels of autocorrelation (Tanir and Booth,
1999; Brown et al., 2005b; Shen and Huang, 2005). Call arrivals may also be heteroscedastic with vari-
ance at least proportional to arrival counts (Taylor, 2008), and overdispersed under a Poisson assumption
having variance per time period typically much larger than its expected value (Jongbloed and Koole, 2001;
Avramidis et al., 2004; Steckley et al., 2005). These properties have implications for various approaches to
modelling and forecasting call arrivals.

148 This subsection was written by Devon K. Barrow.

142
 Lambdas = (1690000,0,0) (39.1,0,3570) (0.666,1.26,797000)

4 5 6 7 8
Observed, Trend

1
Daily seas.

xTime
−2
Weekly seas.

0.0

xTime
0.4−1.0
Random

−0.2

xTime
7 −0.8
8
Fit/Forecast

xTime
5
4

Nov 20 Nov 30 Dec 10 Dec 20

xTime
Figure 15: STR decomposition of the log transformed training data and the forecasts for the traffic flow data.
2500

100
2000
1500

Error %

50
Rate

1000

0
500
0

Dec 18 Dec 20 Dec 22 Dec 18 Dec 20 Dec 22

Figure 16: Left: forecast (red) and the test data (black); Right: the prediction error over time for the traffic flow data.

Figure 17: The staffing decision process in call centres.

143
 The first family of methods are time series methods requiring no distributional assumptions. Early
studies employed auto regressive moving average (ARMA) models (Andrews and Cunningham, 1995;
Tandberg et al., 1995; Xu, 1999; Antipov and Meade, 2002), exponential smoothing (Bianchi et al., 1993,
1998), fast Fourier transforms (Lewis et al., 2003), and regression (Tych et al., 2002). The first meth-
ods capable of capturing multiple seasonality were evaluated by Taylor (2008) and included double sea-
sonal exponential smoothing (Taylor, 2003b) and multiplicative double seasonal ARMA (SARMA). Since
then several advanced time series methods have been developed and evaluated (Taylor, 2010; De Livera
et al., 2011; Taylor and Snyder, 2012), including artificial neural networks (Millán-Ruiz and Hidalgo, 2013;
Pacheco et al., 2009; Li et al., 2011) and models for density forecasting (Taylor, 2012).
Another family of models relies on the assumption of a time-inhomogeneous Poisson process adopting
fixed (Jongbloed and Koole, 2001; Brown et al., 2005b; Shen and Huang, 2008a; Taylor, 2012) and mixed
modelling (Avramidis et al., 2004; Aldor-Noiman et al., 2009; Ibrahim and L’Ecuyer, 2013) approaches to
account for the overdispersed nature of the data, and in some cases, interday and intraday dependence.
The works by Weinberg et al. (2007) and Soyer and Tarimcilar (2008) model call volumes from a
Bayesian point of view. Other Bayesian inspired approaches have been adopted mainly for estimating var-
ious model parameters, but also allowing for intraday updates of forecasts (Landon et al., 2010; Aktekin
and Soyer, 2011).
A further class of approach addresses the dimensionality challenge related to high frequency call data
using Singular Value Decomposition (SVD). Shen and Huang (2005) and Shen and Huang (2008a) use the
same technique to achieve dimensionality reduction of arrival data, and to create a forecasting model that
provides both interday forecasts of call volume, and an intraday updating mechanism. Several further
studies have extended the basic SVD approach to realise further modelling innovations, for example, to
forecast call arrival rate profiles and generate smooth arrival rate curves (Shen et al., 2007; Shen and Huang,
2008b; Shen, 2009). A more comprehensive coverage of different forecasting approaches for call arrival rate
and volume can be found in a recent review paper by Ibrahim et al. (2016).

3.8.5. Elections forecasting149

With the exception of weather forecasts, there are few forecasts which have as much public exposure as
election forecasts. They are frequently published by mass media, with their number and disclosure reach-
ing a frenzy as the Election Day approaches. This explains the significant amount of methods, approaches
and procedures proposed and the paramount role these forecasts play in shaping people’s confidence in
(soft/social) methods of forecasting.
The problem escalates because, regardless whether the goal of the election forecast is an attempt to
ascertain the winner in two-choice elections (e.g., a referendum or a Presidential election) or to reach
estimates within the margins of error in Parliamentary systems, the knowledge of the forecasts influences
electors’ choices (Pavı́a et al., 2019). Election forecasts not only affect voters but also political parties,
campaign organizations and (international) investors, who are also watchful of their evolution.
Scientific approaches to election forecasting include polls, information (stock) markets and statistical
models. They can also be sorted by when they are performed; and new methods, such as social media
surveillance, are also emerging (Huberty, 2015; Ceron et al., 2016). Probabilistic (representative) polls

149 This subsection was written by Jose M. Pavı́a.

144
are the most commonly used instrument to gauge public opinions. The progressive higher impact of
non-sampling errors (coverage issues, non-response bias, measurement error: Biemer, 2010) is, however,
severely testing this approach. Despite this, as Kennedy et al. (2017) show in a recent study covering 86
countries and more than 500 elections, polls are still powerful and robust predictors of election outcomes
after adjustments (see, also, Jennings et al., 2020). The increasing need of post-sampling adjustments of
probabilistic samples has led to a resurgence of interest in non-probabilistic polls (Pavı́a and Larraz, 2012;
Wang et al., 2015b; Elliott and Valliant, 2017), abandoned in favour of probabilistic sampling in 1936, when
Gallup forecasted Roosevelt’s triumph over Landon using a small representative sample despite Literacy
Digest failing to do so with a sample of near 2.5 million responses (Squire, 1988).
A person knows far more than just her/his voting intention (Rothschild, 2009) and when s/he makes
a bet, the rationality of her/his prediction is reinforced because s/he wants to win. Expectation polls
try to exloit the first issue (Graefe, 2014), while prediction markets, as efficient aggregators of informa-
tion, exploit both these issues to yield election forecasts. Several studies have proven the performance of
these approaches (Wolfers and Zitzewitz, 2004; Berg et al., 2008; Erikson and Wlezien, 2012; Williams and
Reade, 2016), even studying their links with opinion polls (Brown et al., 2019). Practice has also developed
econometric models (Fair, 1978) that exploit structural information available months before the election
(e.g., the evolution of the economy or the incumbent popularity). Lewis-Beck has had great success in
publishing dozens of papers using this approach (see, e.g., Lewis-Beck, 2005).
Special mention also goes to Election-Day forecasting strategies, which have been systematically com-
missioned since the 1950s (Mitofsky, 1991). Exit (and entrance) polls (Pavı́a, 2010; Klofstad and Bishin,
2012), quick-counts (Pavı́a-Miralles and Larraz-Iribas, 2008), and statistical models (Moshman, 1964; Bernardo,
1984; Pavı́a-Miralles, 2005) have been used to anticipate outcomes on Election Day. Some of these strate-
gies (mainly random quick-counts) can be also employed as auditing tools to disclose manipulation and
fraud in weak democracies (Scheuren and Alvey, 2008).

3.8.6. Sports forecasting150

Forecasting is inherent to sport. First in this section, we look at forecast competitions in sport, and
following this we consider the role forecasts play in sporting outcomes.
Forecast competitions are common; see section 3.8.13. Sport provides a range of forecast competitions,
perhaps most notably the competition between bookmakers and their customers – betting. A bet is a
contingent contract, a contract whose payout is conditional on specified future events occurring. Bets
occur fundamentally because two agents disagree about the likelihood of that event occurring, and hence
it is a forecast.
Bookmakers have been extensively analysed as forecasters; Forrest et al. (2005) evaluated biases in the
forecasts implied by bookmaker odds over a period where the betting industry became more competitive,
and found that relative to expert forecasts, bookmaker forecasts improved.
With the internet age, prediction markets have emerged, financial exchanges where willing participants
can buy and sell contingent contracts. In theory, such decentralised market structures ought to provide the
most efficient prices and hence efficient forecasts (Nordhaus, 1987). A range of papers have tested this in

150 This subsection was written by J. James Reade.

145
the sporting context (Gil and Levitt, 2007; Croxson and Reade, 2014; Angelini and De Angelis, 2019), with
conclusions tending towards a lack of efficiency.
Judgemental forecasts by experts are commonplace too; traditionally in newspapers, but more recently
on television and online. Reade et al. (2020) evaluate forecasts of scorelines from two such experts against
bookmaker prices, a statistical model, and the forecasts from users of an online forecasting competition.
Singleton et al. (2019) find that when forecasters in the same competition revise their forecasts, their fore-
cast performance worsens. This forecasting competition is also analysed by Butler et al. (2020) and Reade
et al. (2020).
Sport is a spectacle, and its commercial success is conditioned on this fact. Hundreds of millions of peo-
ple globally watch events like the Olympics and the FIFA World Cup – but such interest is conditioned on
anticipation, a forecast that something interesting will happen. A superstar is going to be performing, the
match will be a close encounter, or it will matter a lot for a bigger outcome (the championship, say). These
are the central tenets of sport economics back to Neale (1964) and Rottenberg (1956), most fundamentally
the ‘uncertainty of outcome hypothesis’.
Cities and countries bid to host large events like the World Cup based on forecasts regarding the impact
of hosting such events. Forecasts that are often inflated for political reasons (Baade and Matheson, 2016).
Equally, franchise-based sports like many North American sports attract forecasts regarding the impact of
a team locating in a city, usually resulting in public subsidies for the construction of venues for teams to
play at (Coates and Humphreys, 1999). Governments invest in sporting development, primarily to achieve
better performances at global events, most notably the Olympics (Bernard and Busse, 2004).
Many sporting events themselves rely on forecasts to function; high jumpers predict what height they
will be able to jump over, and free diving contestants must state the depth they will dive to. Less formally,
teams will set themselves goals: to win matches, to win competitions, to avoid the ‘wooden spoon’. Here,
forecast outcomes are influenced by the teams, and competitors, taking part in competitions and, as such,
are perhaps less commonly thought of as genuine forecasts.

3.8.7. Forecasting for megaprojects151

Megaprojects are significant activities characterised by a multi-organisation structure, which produces
highly visible infrastructure or asset with very crucial social impacts (Aaltonen, 2011). Megaprojects are
complex, require huge capital investment, several stakeholders are identified and, usually a vast number
of communities and the public are the receivers of the project’s benefits. There is a need megaprojects
especially those that deliver social and economic goods and create economic growth (Flyvbjerg et al., 2003).
Typical features of megaprojects include some or all the following: (i) delivering a substantial piece of
physical infrastructure with a life expectancy that spans across decades, (ii) main contractor or group of
contractors are privately owned and financed, and (iii) the contractor could retain an ownership stake in
the project and the client is often a government or public sector organisation (Sanderson, 2012).
However, megaprojects are heavily laced with extreme human and technical complexities making their
delivery and implementation difficult and often unsuccessful (The RFE Working Group Report, 2015; Mer-
row et al., 1988). This is largely due to the challenge of managing megaprojects including extreme com-
plexity, increased risk, tight budget and deadlines, lofty ideals (Fiori and Kovaka, 2005). Due to the pos-

151 This subsection was written by Konstantia Litsiou.

146
sibility and consequences of megaproject failure (Mišić and Radujković, 2015), forecasting the outcomes
of megaprojects is becoming of growing importance. In particular, it is crucial to identify and assess the
risks and uncertainties as well as other factors that contribute to disappointing outcomes of megaprojects
in order to mitigate them (Flyvbjerg et al., 2003; Miller and Lessard, 2007).
Literature review in forecasting in megaprojects are scarce. However, there are a few themes that have
emerged in the extant literature as characteristics of megaprojects that should be skilfully managed to
provide a guideline for the successful planning and construction of megaprojects (Fiori and Kovaka, 2005;
Flyvbjerg, 2007; Sanderson, 2012). Turner and Zolin (2012) even claim that we cannot even properly
define what success is. They argue that we need to reliable scales in order to predict multiple perspectives
by multiple stakeholders over multiple time frames — so definitely a very difficult long term problem.
This could be done via a set of leading performance indicators that will enable managers of Megaprojects
to forecast during project execution how various stakeholders will perceive success months or even years
into the operation. At the very early stages of a project’s lifecycle, a number of decisions must been taken
and are of a great importance for the performance and successful deliverables/outcomes. Flyvbjerg (2007)
stress the importance of the front-end considerations particularly for Megaprojects Failure to account for
unforeseen events frequently lead to cost overruns.
Litsiou et al. (2019) suggest that forecasting the success of megaprojects is particularly a challenging
and critical task due to the characteristics of such projects. Megaproject stakeholders typically implement
impact assessments and/or cost benefit Analysis tools (Litsiou et al., 2019). As Makridakis et al. (2010) sug-
gested, judgemental forecasting is suitable where quantitative data is limited, and the level of uncertainty
is very high; elements that we find in megaprojects. By comparing the performance of three judgemen-
tal methods, unaided judgement, semi-structured analogies (sSA), and interaction groups (IG), used by a
group of 69 semi-experts, Litsiou et al. (2019) found that, the use of sSA outperforms unaided judgment
in forecasting performance. The difference is amplified further when pooling of analogies through IG is
introduced.

3.8.8. Competing products152

Based on IMS-Health quarterly number of cimetidine and ranitidine packages sold in Italy, the CRCD
model was tested to evaluate a diachronic competition that produced substitution. Cimetidine is a his-
tamine antagonist that inhibits the production of stomach acid and was introduced by Smith, Kline &
French in 1976. Ranitidine is an alternative active principle introduced by Glaxo in 1981 and was found
to have far-improved tolerability and a longer-lasting action. The main effect in delayed competition is
that the first compound spread fast but was suddenly outperformed by the new one principle that mod-
ified its stand-alone regime. Guseo and Mortarino (2012) give some statistical comparisons with the re-
stricted model Krishnan-Bass-Kummar Diachronic model (KBKD) by Krishnan et al. (2000). Previous
results were improved with the UCRCD model in Guseo and Mortarino (2014) by considering a decom-
position of word-of-mouth (WOM) effects in two parts: within-brand and cross-brand contributions. The
new active compound exploited a large cross-brand WOM and a positive within-brand effect. After the
start of competition, cimetidine experienced a negative WOM effect from its own adopters and benefited

152 This subsection was written by Renato Guseo.

147
from the increase of the category’s market potential driven by the antagonist. Forecasting is more realistic
with the UCRCD approach.
Restricted and unrestricted UCRCD models were applied in Germany by Guidolin and Guseo (2016) to
the competition between nuclear power technologies and renewable energy technologies (wind and solar)
in electricity production. Due to the ‘Energiewende’ policy started around 2000, the substitution effect,
induced by competition, is confirmed by the electricity production data provided by BP153 . An advance is
proposed in Furlan et al. (2020) with three competitors and exogenous control functions obtaining direct
inferences that provide a deeper analysis and forecasting improvements in energy transition context.
Previous mentioned intersections between Lotka-Volterra approach and diffusion of innovations com-
petition models suggested a more modulated access to the residual carrying capacity. The Lotka-Volterra
with churn model (LVch) by Guidolin and Guseo (2015) represents ‘churn effects’ preserving within and
cross-brand effects in a synchronic context. LVch includes unrestricted UCRCD model.
An application of LVch model is discussed with reference to the competition/substitution between com-
pact cassettes and compact discs for pre-recorded music in the US market. Obtained results of LVch out-
perform restricted and unrestricted UCRCD analyses. In this context the residual market is not perfectly
accessible to both competitors and this fact, combined with WOM components, allows for better interpre-
tation and forecasting.
A special application of the LVch model, Lotka-Volterra with asymmetric churn (LVac), is proposed in
Guidolin and Guseo (2020). It is based on a statistically oriented reduction: The late entrant behaves as a
standard Bass (1969) model that modifies the dynamics and the evolution of the first entrant in a partially
overlapped market. The case study is offered by a special form of competition where the iPhone produced
an inverse cannibalisation of the iPad. The former suffered a local negative interaction with some benefits:
A long-lasting life cycle and a larger market size induced by the iPad.
A limitation in models for diachronic competition relate to high number of rivals, implying complex
parametric representations with respect to the observed information. A second limitation, but also an
opportunity, is the conditional nature of forecasting if the processes partially depend upon exogenous
control functions (new policy regulations, new radical innovations, regular and promotional prices, etc.).
These tools may be used to simulate the effect of strategic interventions, but a lack of knowledge of such
future policies may affect prediction.

3.8.9. Safety and security154

Safety and security are among the most pressing and complex tasks of modern time as challenges
such as extreme environmental conditions, system faults and malfunctions, fraud activities and disease
outbreaks can cause significant harm to valuable lives and assets if not detected and treated quickly (Ben-
Gal, 2005).
The problem of anomaly detection in environmental sensors are twofold: (i) identifying data issues that
make the data unreliable and untrustworthy (Hand, 2009; Sobhani et al., 2020), and (ii) identifying real
unexpected events such as extreme weather conditions, water contamination. Early detection of anomalies
in data will limit the use of corrupted data in subsequent forecasting applications such as water-quality

153 https://www.bp.com/en/global/corporate/energy-economics/statistical-review-of-world-energy.html (Accessed: 2020-09-01)

154 This subsection was written by Priyanga Dilini Talagala.

148
monitoring and early warning systems (Storey et al., 2011), predicting algal bloom outbreaks leading to fish
kill events and potential human health impacts, forecasting water level and currents (Archer et al., 2003;
Glasgow et al., 2004). A winning team of the Global Energy Forecasting Competition 2014 (GEFCom2014)
also incorporated a model-based approach for anomaly detection and data cleansing (Xie and Hong, 2016).
Talagala et al. (2019) combine data features and an approach based on extreme value theory to identify
anomalies in water quality data. Hong et al. (2010) also discussed the data quality issues associated with
the historical weather data and their impact on short term load forecasting.
With rapid advances in data collection technology, cybersecurity has become a major concern in the
modern world (Singer and Friedman, 2014). Among various cybersecurity issues, data integrity attacks,
where unauthorised parties access protected or confidential data and inject misinformation, are of great
interest to the forecasting community as data quality has a direct impact on forecast accuracy. These mis-
information can cause over-forecasts that demand unnecessary expenses for the upgrade and maintenance
of infrastructure (Luo et al., 2018a). According to Sobhani et al. (2020), a 1% improvement in forecast ac-
curacy can help a large power company to save millions of dollars. Luo et al. (2018a) and Luo et al. (2018b)
address the problem of load forecasting under data integrity attacks using model based approaches. In in-
trusion detection systems, on the other hand, anomalies in system events might indicate a possible breach
of security (Warrender et al., 1999). The feature based approach proposed by Talagala et al. (2020b) has
the ability to deal with such data challenges.
Efficient forecasting models for early detection of disease outbreaks are a must for modern surveillance.
Outbreaks of interest include bioterrorist-attacks, laboratory releases, naturally occurring epidemics such
as the coronavirus pandemic and the avian influenza outbreak. Analysis of the surveillance data in a
timely manner is essential for rapid epidemiological response and thereby reduce morbidity and mortality
(Lotze and Shmueli, 2009). However, early detection of disease outbreaks are challenging as different
diseases cause different signals and patterns and they are clearly visible mostly during the later stage of an
epidemic (Wagner et al., 2001). According to Lotze and Shmueli (2009), most existing temporal detection
methods for bio-surveillance use forecasting component followed by a residual analysis as in (Hutwagner
et al., 2003; Lombardo et al., 2004).

3.8.10. The forecastability of agricultural time series155

The forecasting of agricultural time series falls under the broader group of forecasting commodities, of
which agricultural and related products are a critical subset. While there has been considerable work in
the econometrics and forecasting literature on common factor models in general there is surprisingly little
work so far on the application of such models for commodities and agricultural time series – and this is so
given that there is considerable literature in the linkage between energy and commodities, including agri-
cultural products, their prices and futures prices, their returns and volatilities. Furthermore, a significant
number of papers is fairly recent which indicates that there are many open avenues of future research on
these topics, and in particular for applied forecasting. The literature on the latter connection can consider
many different aspects in modelling as we illustrate below. We can identify two literature strands, a much
larger one on the various connections of energy with commodities and the agricultural sector (and in this

155 This subsection was written by Dimitrios Thomakos.

149
strand we include forecasting agricultural series) and a smaller one that explores the issue of common
factors.
An early reference of the impact of energy on the agricultural sector is Tewari (1990) and then after
a decade we find Gohin and Chantret (2010) on the long-run impact of energy prices on global agricul-
tural markets. Byrne et al. (2013) is an early reference for co-movement of commodity prices followed
by Daskalaki et al. (2014) on common factors of commodity future returns and then a very recent paper
from Alquist et al. (2020) who link global economic activity with commodity price co-movement. The im-
pact of energy shocks on US agricultural productivity was investigated by Wang and McPhail (2014) while
Koirala et al. (2015) explore the non-linear correlations of energy and agricultural prices with Albulescu
et al. (2020) exploring the latter issue further, the last two papers using copulas. Xiong et al. (2015) is
an early reference of forecasting agricultural commodity prices while Kyriazi et al. (2019), Wang et al.
(2019a), and Li et al. (2020d) consider three novel and completely different approaches on forecasting
agricultural prices and agricultural futures returns. López Cabrera and Schulz (2016) explore volatility
linkages between energy and agricultural commodity prices and then Tian et al. (2017) start a mini-stream
on volatility forecasting on agricultural series followed among others by the work of Luo et al. (2019) and of
Degiannakis et al. (2020). de Nicola et al. (2016) examine the co-movement of energy and agricultural re-
turns while Kagraoka (2016) and Lübbers and Posch (2016) examine common factors in commodity prices.
Wei Su et al. (2019) and Pal and Mitra (2019) both investigate the linkages of crude oil and agricultural
prices. Finally, Tiwari et al. (2020) examine the time-frequency causality between various commodities,
including agricultural and metals.
There is clearly room for a number of applications in the context of this recent research, such along the
lines of further identifying and then using common factors in constructing forecasting models, exploring
the impact of the COVID-19 crisis in agricultural production or that of climate changes on agricultural
prices.

3.8.11. Forecasting espresso coffee quality156

Espresso coffee is among the most popular beverages, and its quality is one of the most discussed and
investigated issues. Besides human-expert panels, electronic noses, and chemical techniques, forecasting
the quality of an espresso by means of data-driven approaches, such as association rule mining, is an
emerging research topic (Apiletti and Pastor, 2020; Apiletti et al., 2020).
In this section, a forecasting model of the espresso quality is presented as the result of a data-driven
analysis applied to a real-world data set of espresso brewing by professional coffee-making machines. Cof-
fee ground size, coffee ground amount, and water pressure have been selected among the most influential
external variables. The ground-truth quality evaluation has been performed for each shot of coffee based
on three well-known quality variables selected by domain experts and measured by specific sensors: the
extraction time, the average flow rate, and the espresso volume. An exhaustive set of more than a thousand
coffees has been produced to train a model able to forecast the effect of non-optimal values on the espresso
quality. For each external variable, different categorical values are considered: ground size can be coarse,
optimal, or fine; ground amount can be high, optimal, or low; brewing water pressure can be high, optimal,
or low.

156 This subsection was written by Daniele Apiletti.

150
 The experimental setting of categorical variables allows us to apply association rule mining, a powerful
data-driven exhaustive and explainable approach (Han et al., 2011; Tan et al., 2005), successfully exploited
in different application contexts (Acquaviva et al., 2015; Di Corso et al., 2018).
Association rule mining parameters minimum support and minimum confidence are set to very low val-
ues, i.e., 0.5% and 10% respectively, so that all possible correlations are extracted. Then, a sorting is
performed so that the most interesting rules are presented and included in the forecasting model: the lift
and the confidence metrics have been used to this aim.
Among the most interesting results, we highlight the following. If the water pressure is low, the amount
of coffee ground is too high, and the grinding is fine, then we can forecast a low-quality coffee due to
excessive percolation time, with a confidence near to 100%. If the amount of coffee ground is low, the
ground is coarse, and the pressure is high, then we can forecast a low-quality coffee due to excessive flow
rate. Furthermore, the coarseness of coffee ground generates an excessive flow rate forecast, despite the
optimal values of dosage and pressure, with a very high confidence (near to 100%).
In most of the cases (83%), the espresso volume is in the optimal-quality range. Espresso machines
are in fact designed and calibrated to extract a pre-determined amount of coffee. Hence, the data-driven
forecasting model almost never provides an espresso-volume quality warning. In most of the cases, the
espresso machines produce an adequate amount of coffee even in non-optimal conditions. It is therefore
extremely important the control of the other quality indexes (percolation time and flow rate) to forecast the
coffee quality, since they strongly impact on the final flavour and body of the espresso despite the optimal
volume.

3.8.12. Dealing with logistic forecasts in practice157

The most fascinating aspect of S curve fitting is the ability to predict from early measurements the
final ceiling. This very fact, however, constitutes also the fundamental weakness and the major criticism of
these forecasts because logistic fits on early data can often accommodate very different values for the final
ceiling. Debecker and Modis (1994) have carried out an extensive simulation study aiming to quantify the
uncertainties on the parameters determined by logistic fits. The study was based on 35,000 S curve fits
on simulated data, smeared by random noise and covering a variety of conditions. The fits were carried
out via a χ2 minimisation technique. The study produced lookup tables and graphs for determining the
uncertainties expected on the three parameters of the logistic function.
A frequent point of misunderstanding and confusion is whether a forecaster should fit an S curve to the
cumulative number or to the number per unit of time. Here the forecaster must exercise wise judgment.
What is the “species” and what is the niche that is being filled? To the frustration of business people there
is no universal answer. When forecasting the sales of a new product it is often clear that one should fit the
cumulative sales because the product’s market niche is expected to eventually fill up. But if we are dealing
with something that is going to stay with us for a long time (e.g., the Internet or a smoking habit), then one
should not fit cumulative numbers. At times this distinction may not be so obvious. For example at the
appearance of COVID-19 many people (often amateurs) began fitting S curves to the cumulative number
of infections. Some of them were rewarded because indeed the diffusion of the virus in some countries
behaved accordingly (Debecker and Modis, 2020). But many were frustrated and tried to “fix” the logistic

157 This subsection was written by Theodore Modis.

151
equation by introducing more parameters, or simply gave up on trying to use logistics with COVID-19.
And yet, there are special situations (e.g., the US), which can be illuminated by logistic fits but on the daily
number of infections, not on the cumulative number. As of August 1, 2020, leaving out the three eastern
states that had gotten things under control, the rest of the US displayed two classic S curve steps followed
by plateaus (see figure 18). The two plateaus reflect the number of infections that American society was
willing to tolerate at the time, as the price to pay for not applying measures to restrict the virus diffusion.
Cases per day
USA without New York, New Jersey, and Connecticut
80,000

70,000 2nd wave

1st wave
60,000
Contributing to the plateau trend
50,000

40,000

30,000

20,000

10,000

0
11-Mar 31-Mar 20-Apr 10-May 30-May 19-Jun 9-Jul 29-Jul

Figure 18: Two logistic-growth steps during the early diffusion of COVID-19 in America (March to July, 2020).

No matter what fitting program one uses, the fitted S curve will flatten toward a ceiling as early and
as low as it is allowed by the constraints of the procedure. As a consequence fitting programs may yield
logistic fits that are often biased toward a low ceiling. Bigger uncertainties on the data points accentuate
this bias by permitting larger margins for the determination of the S curve parameters. To compensate for
this bias the user must explore several fits with different weights on the data points during the calculation
of the χ2 . He or she should then keep the answer that gives the highest ceiling for the S curve (most often
obtained by weighting more heavily the recent historical data points). Of course, this must be done with
good justification; here again the forecaster must exercise wise judgment.

3.8.13. Forecasting competitions158

Forecasting competitions provide a “playground” for academics, data scientists, students, practition-
ers, and software developers to compare the forecasting performance of their methods and approaches
against others. Organisers of forecasting competitions test the performance of the participants’ submis-
sions against some hidden data, usually the last window of observations for each series. The benefits from
forecasting competitions are multifold. Forecasting competitions (i) motivate the development of innova-
tive forecasting solutions, (ii) provide a deeper understanding of the conditions that some methods work
and others fail, (iii) promote knowledge dissemination, (iv) provide a much-needed, explicit link between

158 This subsection was written by Fotios Petropoulos.

152
theory and practice, and (v) leave as a legacy usable and well-defined data sets. Participation in forecasting
competitions is sometimes incentivised by monetary prizes. However, the stakes are usually much higher,
including reputational benefits.
The most famous forecasting competitions are the ones organised by Spyros Makridakis. Initially, the
research question focused on the relative performance of simple versus complex forecast. M and M3 com-
petitions (Makridakis et al., 1982; Makridakis and Hibon, 2000) empirically showed that simple methods
(such as exponential smoothing) are equally good compared to other more complex methods and models
(such as ARIMA and neural networks) in point-forecast accuracy – if not better. Moreover, the early Makri-
dakis competitions showed the importance of forecast combinations in increasing predictive accuracy. For
example, the winner of the M3 competition was the Theta method, a simple statistical method that in-
volved the combination of linear regression and simple exponential smoothing forecasts (Assimakopoulos
and Nikolopoulos, 2000).
The M4 competition (Makridakis et al., 2020c) challenged researchers and practitioners alike with
a task of producing point forecasts and prediction intervals for 100 thousand time series of varied fre-
quencies. This time, the main hypothesis focused on the ability of machine learning and neural network
approaches in the task of time series forecasting. Machine learning approaches that focused on each se-
ries independently performed poorly against statistical benchmarks, such as Theta, Damped exponential
smoothing or simple averages of exponential smoothing models. However, the best two performing sub-
missions in the M4 competition (Smyl, 2020; Montero-Manso et al., 2020) used neural network and ma-
chine learning algorithms towards utilising cross-learning. So, the main learning outcome from the M4
competition is that, if utilised properly, machine learning can increase the forecasting performance. Simi-
larly to previous competitions, M4 demonstrated again the usefulness of combining across forecasts, with
five out of the top six submissions offering a different implementation of forecast combinations.
Several other forecasting competitions focused on specific contexts and applications. For example, M2
competition (Makridakis et al., 1993) suggested that the benefits from additional information (domain ex-
pertise) are limited. The tourism forecasting competition (Athanasopoulos et al., 2011) also showed that
exogenous variables do not add value, while naive forecasts perform very well on a yearly frequency. The
NN3 competition (Crone et al., 2011) confirmed the superior performance of statistical methods, but noted
that neural network approaches are closing the distance. Tao Hong’s series of energy competitions (Hong
et al., 2014, 2016, 2019) demonstrated best practices for load, price, solar, and wind forecasting, with
extensions to probabilistic and hierarchical forecasts. Finally, many companies have hosted forecasting
challenges through the Kaggle platform. Bojer and Meldgaard (2020) reviewed the Kaggle competitions
over the last five years, and concluded that access to hierarchical information, cross-learning, feature en-
gineering, and combinations (ensembles) can lead to increased forecasting performance, outperforming
traditional statistical methods. These insights can be a forerunner to the results of the M5 competition.

3.9. The future of forecasting practice159

Plus ça change, plus c’est la même chose.

Jean-Baptiste Karr (1849)

159 This subsection was written by Len Tashman.

153
 It would be a more straightforward task to make predictions about the future of forecasting practice if
we had a better grasp of the present state of forecasting practice. For that matter, we lack even a common
definition of forecasting practice. In a recent article, Makridakis et al. (2020a) lamented the failure of truly
notable advances in forecasting methodologies, systems, and processes during the past decades to convince
many businesses to adopt systematic forecasting procedures, leaving a wide swath of commerce under the
guidance of ad hoc judgment and intuition. At the other extreme, we see companies with implementations
that combine state-of-the-art methodology with sophisticated accommodations of computing time and
costs, as well as consideration of the requirements and capabilities of a diverse group of stakeholders
(Yelland et al., 2019). So, it is not hyperbole to state that business forecasting practices are all over the
place. What surely is hyperbole, however, are the ubiquitous claims of software providers about their
products accurately forecasting sales, reducing costs, integrating functions, and elevating the bottom line
(Makridakis et al., 2020a; Sorensen, 2020). For this section, we grilled a dozen practitioners and thought
leaders (“the Group”) about developments playing out in the next decade of forecasting practice, and have
categorised their responses:

• Nature of forecasting challenges;

• Changes in the forecasting toolbox;
• Evolution in forecasting processes such as integration of planning functions;
• Expectations of forecasters; and
• Scepticism about real change.

Forecasting Challenges: Focusing on operations, the Group sees demand forecasting becoming ever more
difficult due to product/channel proliferation, shorter lead times, shorter product histories, and spikes in
major disruptions.

• Operational forecasts will have shorter forecast horizons to increase strategic agility required by
business to compete, sustain, and survive.
• New models will need to incorporate supply-chain disruption. Demand chains will need to be
restarted, shortening historical data sets and making traditional models less viable due to limited
history.
• Lead times will decrease as companies see the problems in having distant suppliers. Longer lead
times make accurate forecasting more difficult.

Forecasting Tool Box: Unsurprisingly, this category received most of the Group’s attention. All predict
greater reliance on AI/ML for automating supply-and-demand planning tasks and for reconciling discrep-
ancies in hierarchical forecasting. Longer-horizon causal forecasting models will be facilitated by big data,
social media, and algorithmic improvements by quantum computing. Post-COVID, we will see a greater
focus on risk management/mitigation. The Cloud will end the era of desktop solutions.

• Quantum computers will improve algorithms used in areas like financial forecasting (e.g., Monte
Carlo simulations), and will change our thinking about forecasting and uncertainty.
• Although social media is a tool for “what’s trending now”, new models will be developed to use
social-media data to predict longer-term behaviour. Step aside Brown (exponential smoothing) and
Bass (diffusion).
154
 • Greater automation of routine tasks (data loading, scrubbing, forecast generation and tuning, etc.)
through AI/ML-powered workflow, configurable limits, and active alerts. More black box under the
hood, but more clarity on the dashboard.
• Greater focus on risk management/mitigation through what-if scenarios, simulations, and proba-
bilistic forecasting.

Forecasting Processes and Functional Integration: Systems will become more integrated, promoting greater
collaboration across functional areas and coordination between forecast teams and those who rely upon
them. Achieving supply-chain resilience will become as important as production efficiency, and new tech-
nology such as Alert and Root Cause Analysis systems will mitigate disruptions.

• S&OP will expand from its home in operations to more fully integrate with other functions such as
finance and performance management, especially in larger multinationals.
• The pandemic has forced firms to consider upping supply-chain resilience. Firms are building in
capacity, inventory, redundancy into operations—somewhat antithetical to the efficiency plays that
forecasting brings to the table.
• Forecasting will be more closely tied to Alert and Root Cause Analysis systems, which identify break-
downs in processes/systems contributing to adverse events, and prevent their recurrence.

Expectations of Forecasters: Agreement was universal that the forecaster’s job description will broaden and
become more demanding, but that technology will allow some redirection of effort from producing fore-
casts to communicating forecasting insights.

• The interest around disease models increases our awareness of the strengths and weaknesses of math-
ematical models. Forecasters may need to become more measured in their claims, or do more to resist
their models being exploited.
• We will see a transformation from demand planner to demand analyst, requiring additional skill sets
including advanced decision making, data and risk analysis, communication, and negotiation.
• Professional forecasters will be rare except in companies where this expertise is valued. Fewer stu-
dents are now educated or interested in statistical modelling, and time is not generally available for
training.
• Forecasters will learn the same lesson as optimisation folks in the 1990s and 2000s: the importance
of understanding the application area—community intelligence.

Scepticism: Many were sceptical about the current enthusiasm for AI/ML methods; disappointed about the
slow adoption of promising new methods into software systems and, in turn, by companies that use these
systems; and pessimistic about the respect given to and influence of forecasters in the company’s decision
making.

• While AI/ML are important additions to the forecaster’s toolbox, they will not automatically solve
forecasting issues. Problems include data hunger, capacity brittleness, dubious input data, fickle
trust by users (Kolassa, 2020c), and model bias.
• Practices in the next decade will look very similar to the present. Not that much has changed in the
last decade, and academic developments are slow to be translated into practice.

155
• Politics, gaming, and the low priority given to forecasting are the prime drivers of practice, thus
limiting interest in adopting new methodologies.
• None of the topical items (AI/ML, big data, demand sensing, new forecasting applications) will have
much of an impact on forecasting practice. Forecasting departments hop from one trend to the other
without making much progress towards better forecasting accuracy.
• Software companies will struggle, despite good offerings. Most companies do not want to invest in
excellent forecasting engines; whatever came with their ERP system is “good enough”.
• Forecasting will continue to suffer from neglect by higher levels of management, particularly when
forecasts are inconveniently contrary to the messages management hopes to convey.

156
4. Forecasting: benefits, practices, value, and limitations160

Mr. Buffett said his advice for the cash left to

his wife was that 10 per cent should go to
short-term government bonds and 90 per cent
into a very low-cost S&P 500 index fund.

In this paper, there are more than 140 sections and subsections written by 74 of some of the best-
known forecasting researchers and practitioners in the world, making it into a selective, encyclopedic
piece covering, into a single source, a great deal of the available knowledge about the theory and practice
of forecasting. But some people argue if there is any value in attempting to predict the future and if
forecasting is any different than fortune telling, given the large numbers of mistaken forecasts made in
the past, including our inability to predict the progression of COVID-19 and its economic and human
consequences? What is, therefore, the usefulness of a paper like the present one when crystal balling is
not possible, and uncertainty reigns? It is the aim of this concluding article to set the record straight,
explaining the benefits and practical value of forecasting while reporting its limitations too.
The Myriad of Forecasts: All planning and the great majority of decisions we make require forecasting.
Deciding what time to get up in the morning, not to be late for work implies a correct prediction of the
commuting time to go to the office. Determining what to study is another decision requiring elaborate pre-
dictions about the demand for future jobs decades away. In the business world, firms must decide/forecast
how many products to manufacture, the price they should be sold, how much money to spend on adver-
tising and promotion, how much and in what type of new technologies to invests and a plethora of other
future-oriented decisions requiring both predictions and assessing their inevitable uncertainty. Whether
we like it or not, we have no choice but making these forecasts to benefit as much as possible from their
value, knowing perfectly well that all predictions are uncertain and many turn out to be wrong.
The Pervasiveness of Uncertainty: Apart from some areas of hard sciences, all other forecasts are un-
certain and must be accompanied with a measure of its magnitude, expressed as a prediction interval, or
as a probability distribution around the most likely forecast. Although the value and usage of forecasts
is clear, that of uncertainty is not. Worse, it becomes an unwelcome source of anxiety whose usefulness
is misunderstood. Executives want to know the exact sales of their firm for next month to set up their
production schedule. Instead, they are given prediction intervals (PIs) around such forecast and told that
most of the time, sales will be within this interval, assuming that the fluctuations follow some distribu-
tional assumptions. They argue that forecasting must decrease, not amplify, uncertainty and that the PIs
are too wide and ‘uninformative’ to be used for making practical business decisions. The trouble is that
these PIs are based on past fluctuations and present the best estimation of future uncertainty, even if they
seem too wide. Worse, empirical research has shown that they are actually too narrow, underestimating
uncertainty often considerably.
Assessing Uncertainty and Dealing with its Implied Risks: Uncertainty entails risks, requiring action to
minimise their negative consequences. There are two kinds of uncertainty that can be illustrated by the

160 This subsection was written by Spyros Makridakis.

157
commuting example. The first relates to fluctuations in the commuting time under normal driving condi-
tions when there are no serious accidents, road works or major snowstorms. Such fluctuations are small
and can be captured by a normal curve that allows to balance the risk of arriving earlier or later than the
desired time. In the opposite case, uncertainty is fat-tailed and hard to estimate, as delays can be sub-
stantial depending upon the seriousness of the accident or that of the snowstorm while the risk of being
early to work is smaller than being late. Moreover, such risk is substantially different when going to the
airport to catch a flight, requiring starting much earlier than the average time it takes to go to the airport
to minimise the risk of missing the flight.
More Accurate Ways of Forecasting and Assessing Uncertainty: Extensive empirical research, including
forecasting competitions, has shown that systematic approaches improve the accuracy of forecasting and
the correct assessment of uncertainty resulting in substantial benefits when compared to ad-hoc judgmen-
tal alternatives (Makridakis et al., 2020a). The biggest advantage of such approaches is their ability to
identify and estimate, in a mathematically optimal manner, past patterns and relationships that are sub-
sequently extrapolated to predict their continuation, avoiding the over optimism and wishful thinking
associated with judgmental approaches. At the same time, it must be clear that the accuracy of the fore-
casts and the correctness of uncertainty will depend on the established patterns/relationship not changing
much during the forecasting horizon.
Using Benchmarks to Evaluate the Value of Forecasting: The accuracy of the forecasts and the correct
assessment of uncertainty must be judged not on their own but in comparison to some simple, readily
available benchmarks. In stock market forecasts, for instance, the accuracy of predictions is compared to
that of today’s price used as the forecast for future periods. Empirical comparisons have shown that such a
benchmark beats the great majority of professional forecasters, hence Buffet’s advice in the epigram for his
wife to invest in a low-cost index fund that selects stocks randomly. In weather forecasting, meteorologists
are judged by the improvement of their forecasts over the naive prediction that tomorrow’s weather will
be the same as today.
Concluding remark: Accepting the advantages and limitations of systematic forecasting methods and
most importantly avoiding any exaggerated expectations of what it can achieve is critical. Such methods
do not possess any prophetic powers, they simply extrapolate established patterns and relationships to
predict the future and assess its uncertainty. Their biggest advantage is their objectivity and ability for
optimal extrapolation. Their biggest disadvantages are: (i) the patterns and the relationships must remain
fairly constant during the forecasting phase for the forecasts to be accurate, and (ii) uncertainty must not
be fat-tailed so it can be measured quantitatively.

158
Disclaimer

The views expressed in this paper are those of the authors and do not necessarily reflect the views of
their affiliated institutions and organisations.

Acknowledgements

Fotios Petropoulos would like to thank all the co-authors of this article for their very enthusiastic re-
sponse and participation in this initiave. He would also like to thank Pierre Pinson for inviting this paper to
be submitted to the International Journal of Forecasting and for forming an advisory board. The constructive
comments and suggestions from this advisory board were vital in improving the paper.
Clara Cordeiro is partially financed by national funds through FCT – Fundação para a Ciência e a
Tecnologia under the project UIDB/00006/2020.
Fernando Luiz Cyrino Oliveira acknowledges the support of the Coordination for the Improvement of
Higher Level Personnel (CAPES) – grant number 001, the Brazilian National Council for Scientific and
Technological Development (CNPq) – grant number 307403/2019-0, and the Carlos Chagas Filho Re-
search Support Foundation of the State of Rio de Janeiro (FAPERJ) – grant numbers 202.673/2018 and
211.086/2019.
Shari De Baets was funded by the FWO Research Foundation Flanders.
Piotr Fiszeder was supported by the National Science Centre project number 2016/21/B/HS4/00662
entitled “Multivariate volatility models - the application of low and high prices”.
Mariangela Guidolin acknowledges the support of the University of Padua, Italy, through the grant
BIRD188753/18.
Yanfei Kang acknowledges the support of the National Natural Science Foundation of China (No.
11701022) and the National Key Research and Development Program (No. 2019YFB1404600).
Stephan Kolassa would like to thank Tilmann Gneiting for some very helpful tips.
Jose M. Pavı́a acknowledges the support of the Spanish Ministry of Science, Innovation and Universities
and the Spanish Agency of Research, co-funded with FEDER funds, grant ECO2017-87245-R, and of Con-
sellerı́a d’Innovació, Universitats, Ciència i Societat Digital, Generalitat Valenciana, grant AICO/2019/053.
Diego J. Pedregal Juan Ramon Trapero Arenas acknowledge the support of the European Regional De-
velopment Fund and Junta de Comunidades de Castilla-La Mancha (JCCM/FEDER, UE) under the project
with reference SBPLY/19/180501/000151 and by the Vicerrectorado de Investigación y Polı́tica Cientı́fica
from UCLM through the research group fund program (PREDILAB; DOCM 26/02/2020 [2020-GRIN-
28770].
David E. Rapach thanks Ilias Filippou and Guofu Zhou for valuable comments.
J. James Reade and Han Lin Shang acknowledge Shixuan Wang for his constructive comments.

159
Appendix A. List of acronyms

ACC: Ant Colony Clustering

ACD: Autoregressive Conditional Duration
ADIDA: Aggregate–Disaggregate Intermittent Demand Approach
ADL: Autoregressive Distributed Lag
ADMM: Alternating Direction Method of Multipliers
AI: Artificial Intelligence
AIC: Akaike’s Information Criterion
ANACONDA: Analysis of National Causes of Death for Action
ANFIS: Adaptive Neuro-Fuzzy Inference System
ANN: Artificial Neural Network
AO: Additive Outlier
AR: AutoRegressive (model)
ARX: AutoRegressive with eXogenous variables (model)
ARCH: AutoRegressive Conditional Heteroskedasticity
ARMA: AutoRegressive-Moving Average (model)
ARIMA: AutoRegressive Integrated Moving Average (model)
ARIMAX: AutoRegressive Integrated Moving Average with eXogenous variables (model)
B&M: Brick and Mortar
BATS: Box-Cox transform, ARMA errors, Trend, and Seasonal components (model)
BEER: Behavioural Equilibrium Exchange Rate
BEKK: Baba-Engle-Kraft-Kroner GARCH
BEKK-HL: BEKK High Low
BIC: Bayesian Information Criterion
BLAST: Building Loads Analysis and System Thermodynamics
BM: Bass Model
BMC: Bootstrap Model Combination
BPNN: Back-Propagation Neural Network
CARGPR: Conditional AutoRegressive Geometric Process Range (model)
CARR: Conditional AutoRegressive Range (model)
CARRS: Conditional AutoRegressive Rogers and Satchell (model)
CDS: Credit Default Swap
CNN: Convolutional Neural Network
COVID-19: Coronavirus disease 2019
CVAR: Cointegrated Vector AutoRegressive (model)
CAViaR: Conditional AutoRegressive Value At Risk
CL: Cross Learning
CPFR: Collaborative Planning, Forecasting, and Replenishment
CRCD: Competition and Regime Change Diachronic (model)
CRPS: Continuous Ranked Probability Score CSR: Complete Subset Regression
CV: Cross-Validation

160
DA: Deterministic Annealing
DC: Distribution Centre
DCC: Dynamic Conditional Correlation
DCC-RGARCH: Range GARCH DCC
DFM: Dynamic Factor Model
DGP: Data Generation Process
DJI: Dow Jones Industrial DM: Diebold-Mariano (test)
DSGE: Dynamic Stochastic General Equilibrium
DSHW: Double Seasonal Holt-Winters
DSTCC-CARR: Double Smooth Transition Conditional Correlation CARR
DT: Delay Time
EEG: ElectroEncephaloGram
EGARCH: Exponential GARCH
EH: Expectations Hypothesis
EMD: Empirical Mode Decomposition
ENet: Elastic Net
ENSO: El Niño Southern Oscillation
ERCOT: Electric Reliability Council of Texas
ES: Expected Shortfall
ESP-r: Environmental Systems Performance – research
ESTAR: Exponential STAR ETS: ExponenTial Smoothing (or Error, Trend, Seasonality)
EVT: Extreme Value Theory
EWMA: Exponentially Weighted Moving Average
FAR: Functional AutoRegressive (model)
FASSTER: Forecasting with Additive Switching of Seasonality, Trend and Exogenous Regressors
FCM: Fuzzy C-Means
FIGARCH: Fractionally Integrated GARCH
FIS: Fuzzy Inference System
FFNN: Feed-Forward Neural Network
FFORMA: Feature-based FORecast Model Averaging
FMCG: Fast Moving Consumer Goods
FPCA: Functional Principal Component Analysis
FRB/EDO: Federal Reserve Board’s Estimated, Dynamic, Optimisation-based (model)
FSS: Forecasting Support System
FVA: Forecast Value Added
GARCH: General AutoRegressive Conditional Heteroscedasticity
GARCH-PARK-R: GARCH PARKinson Range
GARCH-TR: GARCH True Range
GB: Givon-Bass
GBM: Generalised Bass Model
GDP: Gross Domestic Product

161
GGM: Gusseo-Guidolin Model (GGM)
GJR-GARCH: Glosten-Jagannathan-Runkle GARCH
GM: Generalised M-estimator
GMM: Generalised Methods of Moments
GPU: Graphics Processing Unit
GRNN: Generalised Regression Neural Network
HAR: Heterogeneous AutoRegressive (model)
HDFS: Hadoop Distributed File System
HMD: Human Mortality Database
HP: Hodrick-Prescott
HPU: House Price Uncertainty
HVAC: Heating, Ventilation, and Air Conditioning (system)
HAR: Heterogeneous AutoRegressive (model)
HQ: Hannan-Quinn
IEA: International Energy Agency
IG: Interaction Groups
iid: independent and identically distributed
IIS: Impulse Indicator Saturation
IO: Innovation Outlier
IT: Information Technology
KBKD: Krishnan-Bass-Kummar Diachronic (model)
KISS: Keep It Simple, Stupid (principle)
kNN: k Nearest Neighbours
KPSS: Kwiatkowski–Phillips–Schmidt–Shin
L-IVaR: Liquidity-adjusted Intraday Value-at-Risk
LASSO: Least Absolute Shrinkage and Selection Operator
LH: Low and High
LLN: Law of Large Numbers
LN-CASS: Logit-Normal Continuous Analogue of the Spike-and-Slab
LS (or LPS): Logarithmic (Predictive) Score (log-score)
LSTAR: Logistic STAR
LSTM: Long Short-Term Memory Networks
LV: Lotka-Volterra model
LVac: Lotka-Volterra with asymmetric churn (model)
LVch: Lotka-Volterra with churn (model)
MAE: Mean Absolute Error
MAPE: Mean Absolute Percentage Error
MASE: Mean Absolute Scaled Error
MFI: Marginal Forecast Interval
MIDAS: MIxed DAta Sampling
MJO: Madden Julian Oscillation

162
ML: Machine Learning
MLP: MultiLayer forward Perceptron
MLR: Multiple Linear Regression
MS VAR: Markov Switching VAR
MSARIMA: Multiple/Multiplicative Seasonal ARIMA
MSC: Multiple Seasonal Cycles
MSE: Mean Squared Error
MSM: Markov Switching Model
MSRB: Markov-Switching Range-Based
MTA: Multiple Temporal Aggregation
NAO: North Atlantic Oscillation
NLS: Nonlinear Least Squares
NLTK: Natural Language Toolkit
NMAE: Normalised Mean Absolute Error
NN: Neural Network
NNAR: Neural Network AutoRegressive
NOB: Non-Overlapping Blocks
NWP: Numerical Weather Prediction
OBR: Office for Budget Responsibility
OB: Overlapping Blocks
OLS: Ordinary Least Squares
OWA: Overall Weighted Average
PAR: Periodic AutoRegressive (model)
PCA: Principal Components Analysis
pdf: probability density function
PdM: Predictive Maintenance
PFEM: Point Forecast Error Measure
PHANN: Physical Hybrid Artificial Neural Network
pHDR: predictive Highest Density Region
PI: Prediction Interval
PIT: Probability Integral Transform
PL: Product Level
PLS: Partial Least Squares
PM: Particulate Matter
POT: Peak Over Threshold
PPP: Purchasing Power Parity
PSO: Particle Swarm Intelligence
PV: PhotoVoltaic
RAF: Royal Air Force (UK)
RB: Range-Based
RB-copula: Range-Based copula

163
RB-DCC: Range-Based DCC
RB-MS-DCC: Range-Based Markov-Switching DCC
RBF: Radial Basis Function
REGARCH: Range-Based Exponential GARCH
RET: Renewable Energy Technology
RGARCH: Range GARCH
RMSE: Root Mean Squared Error
RNN: Recurrent Neural Network
RR-HGADCC: Return and Range Heterogeneous General Asymmetric DCC
RTV: Real Time Vintage
SA: Structured Analogies
SARIMA: Seasonal AutoRegressive Integrated Moving Average (model)
SARIMAX: Seasonal AutoRegressive Integrated Moving Average with eXogenous variables
SARMA: Seasonal AutoRegressive Moving Average (model)
SARMAX: Seasonal AutoRegressive Moving Average with eXogenous variables
SBA: Syntetos-Boylan Approximation
SBC: Syntetos-Boylan-Croston (classification)
SEATS: Seasonal Extraction in ARIMA Time Series
SES: Simple (or Single) Exponential Smoothing
SETAR: Self-Exciting Threshold AutoRegressive (model)
SFI: Simultaneous Forecast Interval
SKU: Stock Keeping Unit
SGD: Stochastic Gradient Descent
SIS: Step Indicator Saturation
SL: Serial number Level
SMA: Simple Moving Average
SOM: Self-Organising Map
SS: State Space
sSA: semi-Structured Analogies
SSARIMA: Several Seasonalities (or State Space) ARIMA
STAR: Smooth Transition AutoRegressive (model)
STARR: Smooth Transition conditional AutoRegressive Range (model)
STL: Seasonal Trend decomposition using Loess
STLF: Short-Term Load Forecasting
STR: Seasonal-Trend decomposition based on Regression
SV: Stochastic Volatility
SVA: Stochastic Value Added
SVD: Singular Value Decomposition
SVM: Support Vector Machine
SWAN: Simulating WAves Nearshore
S&OP: Sales and Operations Planning

164
S&P: Standard & Poor’s
TAR: Threshold AutoRegressive (model)
TARMA: Threshold AutoRegressive Moving Average (model) TARMASE: Threshold AutoRegressive Mov-
ing Average (model)
TARR: Range-Based Threshold conditional AutoRegressive (model)
TBATS: Exponential Smoothing state space model with Box-Cox transformation, ARMA errors, Trend and
Seasonal components
TGARCH: Threshold GARCH
TMA: Threshold Moving Average (model)
TPU: Tensor Processing Unit
TRAMO: Time series Regression with ARIMA noise, Missing values and Outliers
TSB: Teunter-Syntetos-Babai (method)
UCRCD: Unbalanced Competition and Regime Change Diachronic (model)
UIP: Uncovered Interest Party
VaR: Value at Risk
VAR: Vector AutoRegressive (model)
VARX: VAR with eXogenous variables (model)
VARMA: Vector AutoRegressive Moving Average (model)
VAT: Value Added Tax
VARIMAX: Vector AutoRegressive Integrated Moving Average with eXogenous variables (model)
VECM (or VEC): Vector Error Correction Model
VEqCM: Vector Equilibrium-Correction Model
WLS: Weighted Least Squares
WNN: Wavelet Neural Network
WOM: Word-Of-Mouth
WW2: World War 1
WW2: World War 2
WW3: World War 3
XGBoost: eXtreme Gradient Boosting

165
 Appendix B. Software

Table A.1: A list of indicative free or open-source packages, libraries, and toolboxes linking to the theory sections of this article. The authors
assume no liability for the software listed below; interested users are strongly advised to read the respective documentations and licences
terms.

Related section Software Package/Library/Toolbox Function(s) Comments

2.2.1. Unit root tests R aTSA adf.test Augmented Dickey-Fuller test.
R tseries kpss.test Kwiatkowski-Phillips-Schmidt-Shin (KPSS) test.
R forecast ndiffs; nsdiffs Estimates the number of (seasonal) differences required to make a given time
series stationary.
2.2.2. Time series decomposition R stats decompose; stl Classical decomposition method (additive and multiplicative), and STL de-
composition method.
R forecast seasadj; seasonal; mstl; msts; Tools for extracting components, and multiple seasonal decomposition
tbats.components methods.
R tsutils decomp; seasplot Classical decomposition method, and functions for seasonal plots.
R stR AutoSTR; STR; heuristicSTR; Seasonal-Trend decomposition based on Regression.
plot.STR; seasadj.STR
R seasonal seas Functions for X-11, SEATS, and X-13-ARIMA-SEATS decomposition meth-
ods.
166

2.2.3. Anomaly detection and time series forecast- R anomalize time decompose; anomalize; A “tidy” workflow for detecting anomalies in data.
ing time recompose
R oddstream find odd streams; Early detection of anomalous series within a large collection of streaming
extract tsfeatures; time series data.
set outlier threshold
R tsoutliers tso; locate.outliers.oloop; re- Detection of outliers in time series such as Innovational outliers, additive
move.outliers outliers, level shifts, temporary changes and seasonal level shifts.
R stray find HDoutliers; Anomaly detection in high dimensional and temporal data.
find threshold; dis-
play HDoutliers
R forecast tsoutliers; tsclean Provides some simple heuristic methods for identifying and correcting out-
liers.
R OutliersO3 OutliersO3; O3plotM; Draws overview of outliers (O3) Plots.
O3plotT; O3prep
R CRAN Task View Anomaly Detection with R Contains a list of R packages that can be used for anomaly detection
(https://github.com/pridiltal/ctv-AnomalyDetection).
2.3.1. Exponential smoothing models R forecast ets; forecast.ets; ses; Functions for simple exponential smoothing and automatic exponential
smoothing modelling.
R smooth es Function for automatic exponential smoothing modelling.
2.3.2. Time-series regression models R stats lm Fitting linear regression models.
R leaps regsubsets Functions for selecting linear regression models.
R relaimpo Relative importance of regressors in linear models.
R MASS stepAIC Choose a model by AIC in a stepwise algorithm.
2.3.3. Theta method and models R forecast thetaf Returns forecasts and prediction intervals for a theta method forecast.
 Related section Software Package/Library/Toolbox Function(s) Comments
2.3.3. Theta method and models (continued) R forecTheta stheta; stm; otm; dstm; dotm Functions for forecasting univariate time series using Theta Models.
R tsutils theta Estimate Theta method.
2.3.4. Autoregressive integrated moving average R forecast auto.arima; Arima; arfima; Functions for fitting and forecasting with ARIMA models.
(ARIMA) models arima.errors; arimaorder
R smooth auto.msarima; auto.ssarima; State-space and multiple seasonalities implementations of ARIMA models.
msarima; ssarima
2.3.5. ARCH/GARCH models R tseries garch Fit GARCH models to time series.
Python PyFlux Time series analysis and prediction tools that focus on
autoregressive methods (ARIMA, ARCH, GARCH, etc)
(https://pyflux.readthedocs.io/en/latest/index.html).
2.3.6. Forecasting for multiple seasonal cycles R smooth msarima; ssarima Functions for forecasting data with multiple seasonal cycles.
R fable model; forecast; fasster; ETS; Forecasting models for tidy time series.
ARIMA; TSLM
R, Python prophet Facebook’s automatic forecasting procedure.
R tidymodels Collection of packages for modelling and machine learning using tidyverse
principles.
R forecast tbats; dshw Functions for forecasting data with multiple seasonal cycles.
R fable.prophet prophet; forecast A tidy R interface to the prophet forecasting procedure using fable
(https://github.com/mitchelloharawild/fable.prophet).
2.3.7. State-space models Matlab SSpace General modelling of linear, non-linear and non-Gaussian State Space sys-
tems.
167

Matlab SSM General modelling of linear, non-linear and non-Gaussian State Space sys-
tems.
Matlab SSMMATLAB A Set of MATLAB Programs for the Statistical Analysis of State Space Mod-
els.
Matlab E4 A MATLAB toolbox for time series analysis in State Space form.
R UComp Automatic identification of Unobserved Components models in State Space
form.
R statespacer State Space modelling, mainly ARIMA and Basic Structural Models.
R smooth Forecasting using single error State Space models.
R bssm Bayesian Inference of Non-Gaussian State Space Models.
R mssm Multivariate State Space models.
R KFAS Kalman Filter and Smoother for Exponential Family State Space Models.
R TSSS Time Series Analysis with State Space Model, based on the methods in Kita-
gawa (1993).
R dlm Bayesian and Likelihood Analysis of Dynamic Linear Models (Gaussian
State Space models).
Python statsmodels statespace Time Series Analysis by State Space Methods.
2.3.8. Markov switching models R MSwM Fitting Markov switching models.
R NHMSAR Non-homogeneous Markov switching autoregressive models.
2.3.9. Threshold models R TAR Bayesian modelling of autoregressive threshold time series models.
R TSA tar; star Functions for threshold models (and general time series analysis).
2.3.11. Forecasting with many variables R gets getsm, getsv, isat, isatvar Package that implements general to specific model selection, indicator satu-
ration, with functionality for forecasting.
 Related section Software Package/Library/Toolbox Function(s) Comments
2.3.11. Forecasting with many variables (continued) R vars Functions and routines for VAR Modelling.
2.3.12. Functional time series models R ftsa ftsm; farforecast; Functional time series analysis.
T stationarity
2.3.13. Bayesian forecasting with copulas R VineCopula Statistical analysis of vine copula models.
Matlab Patton’s copula toolbox Collection of Matlab function on copulas for financial time series by Andrew
Patton (http://public.econ.duke.edu/ãp172/code.html).
R cdcopula Covariate-dependent copula models (https://github.com/feng-
li/cdcopula).
R FactorCopula Factor Copula Models for Mixed Continuous and Discrete Data.
2.3.14. Forecasting with DSGE models R BMR forecast Bayesian Macroeconometrics in R (BMR) is a package for estimating and
forecasting Bayesian VAR and DSGE.
Matlab/GNU Oc- Dynare Software platform for solving, estimating, and making forecasts with DSGE
tave (https://www.dynare.org/download/).
2.3.16. Robust Estimation and Forecasting R gets isat Function for running impulse and step indicator saturation.
2.3.19. Innovation diffusion models R DIMORA Estimation of Bass Model, generalised Bass Model, GGM.
R diffusion diffusion Various diffusion models to forecast new product growth. Currently the
package contains Bass, Gompertz and Gamma/Shifted Gompertz curves.
2.3.20. The natural law of growth in competition R LS2Wstat scurve An S curve function between two constant values.
2.3.22. Estimation and representation of uncer- R hdrcde cde Conditional kernel density estimation to produce marginal distributions
tainty (uncertainty forecasts).
R gamlss gamlss Semi-parametric models for uncertainty forecasting.
168

R gamboostLSS mboostLSS Semi-parametric component-wise gradient boosting models for uncertainty

forecasting.
Python scikit-learn GradientBoostingRegressor; Machine learning models (gradient boosting trees and random forests) for
RandomForestQuantileRe- quantile forecasting.
gressor
R quantreg rq; lprq; nlqr Estimation and inference methods for models of conditional quantiles.
2.3.22. Estimation and representation of uncer- R EnvStats FcnsByCatPredInts; pointwise Functions for computing prediction intervals and simultaneous prediction
tainty (continued) intervals.
R rmgarch dccfit-methods; dccforecast- Multivariate GARCH Models (e.g., forecasting covariance matrix).
methods
2.3.23. Forecasting under fat tails R FatTailsR Functions for Kiener distributions and fat tails.
2.4.1. Leading indicators and Granger causality R lmtest grangertest Test for Granger causality.
2.4.3. Variable Selection R glmnet Generalised linear models with Lasso or elastic net regularisation.
2.4.4. Model Selection R gets Functions for automatic general to specific model selection.
2.4.5. Cross-validation for time-series data R forecast CVar k-fold Cross-Validation applied to an autoregressive model.
2.5.1. Forecast combination: a brief review of sta- R forecastHybrid Functions for ensemble time series forecasts.
tistical approaches
2.6.2. Forecasting on distributed systems Database InfluxDB Scalable datastore for metrics, events, and real-time analytics
(https://www.influxdata.com/time-series-database/).
Database OpenTSDB A scalable, distributed Time Series Database (http://opentsdb.net/).
Database RRDtool A program for easily maintaining a database of time-series data
(https://oss.oetiker.ch/rrdtool/).
 Related section Software Package/Library/Toolbox Function(s) Comments
2.6.2. Forecasting on distributed systems (contin- Database Timely A time series database application that provides secure access to time series
ued) data (https://code.nsa.gov/timely/).
Python, R & darima Implementations of distributed ARIMA models on Spark platform
Spark (https://github.com/xqnwang/darima) .
2.6.3. Agent-based models R SpaDES Spatially explicit discrete event simulation models.
2.6.4. Feature-based time series forecasting R tsfeatures tsfeatures Methods for extracting various features from time series data.
Python tsfresh Calculates a large number of time series characteristics, the so called fea-
tures. Further the package contains methods to evaluate the explaining
power and importance of such characteristics for regression or classification
tasks.
Matlab hctsa Code framework that enables the extraction of thousands of time-series fea-
tures from a time series (or a time-series dataset). It also provides a range of
tools for visualising and analysing the resulting time-series feature matrix
(https://github.com/benfulcher/hctsa).
Python pyopy Python binding for hctsa.
R fforma Tools for forecasting using a model combination approach. It can be used
for model averaging or model selection. It works by training a ‘classifier’
that learns to select/combine different forecast models.
R gratis Efficient algorithms for generating time series with diverse and controllable
characteristics, which can be used as the training data in feature-based time
series forecasting.
169

R seer Implementations of a novel framework for forecast model selection using

time series features; FFORMS (Feature-based FORecast Model Selection).
2.6.5. Forecasting with bootstrap R bootstrap Various bootstrapping functions.
2.6.6. Bagging for time series forecasting R forecast baggedETS Returns forecasts and other information for bagged ETS models.
R tshacks baggedClusterETS; treat- Returns forecasts for bagged Cluster and Treated ETS models
edETS (https://github.com/tiagomendesdantas/tshacks).
2.6.7. Neural Networks Python, MXNet & GluonTS Framework for building deep learning based models including a
PyTorch number of pre-built models such as feed-forward neural networks
(https://github.com/awslabs/gluon-ts).
R nnfor mlp; elm Time Series Forecasting with Neural Networks.
Python neural prophet Reimplementation of prophet in PyTorch, and extensions to it
(https://github.com/ourownstory/neural prophet).
Python RNNs for forecasting in Tensorflow (https://github.com/HansikaPH/time-
series-forecasting).
R ANN2 neuralnetwork Artificial Neural Networks.
R nnet nnet Feed-Forward Neural Networks and Multinomial Log-Linear Models.
R forecast nnetar Auto-regressive Neural Network for time series forecasting.
2.6.8. Deep Probabilistic models Python, MXNet & GluonTS Framework for building deep learning based models including a num-
PyTorch ber of pre-built models such as DeepAR, DeepState models and NBEATS
(https://github.com/awslabs/gluon-ts).
Python, PyTorch PyTorchTS Clone of GluonTS in PyTorch (https://github.com/zalandoresearch/pytorch-
ts).
 Related section Software Package/Library/Toolbox Function(s) Comments
2.6.9. Machine learning R RSNNS mlp; rbf; dlvq; elman; jordan; Neural Networks using the Stuttgart Neural Network Simulator (SNNS).
som
R rpart rpart; prune Recursive partitioning and regression trees.
R caret Classification and regression training.
R e1071 svm Misc ML functions of the Department of Statistics, Probability Theory
Group.
R kernlab gausspr Gaussian processes for regression and classification.
R brnn brnn Bayesian Regularisation for Feed-Forward Neural Networks.
R grnn grnn General regression neural network.
R randomForest randomForest Breiman and Cutler’s Random Forests for Classification and Regression.
R gbm gbm Generalised Boosted regression models.
R neuralnet neuralnet Training of simple Neural Networks.
Python Tensorflow A framework, developed by Google, offering tools for designing, building,
and deploying ML models (https://tensorflow.org/api docs/python/tf).
Python Keras API A deep learning API built on top of Tensorflow. It provides high level blocks
for building and training NN models (https://keras.io/).
R Tensorflow R Interface to Tensorflow (https://tensorflow.rstudio.com/).
R deepnet Deep learning toolkit.
R h2o R Interface for the ’H2O’ Scalable Machine Learning Platform.
2.6.9. Machine learning (continued) R Apache MXNet A flexible library for deep learning (https://mxnet.apache.org/versions/1.7.0/).
Python scikit-learn Ordinary Least Squares, Ridge regression, Lasso, Bayesian Regression, Gen-
170

eralized Linear Regression, Stochastic Gradient Descent and Polynomial

regression, Support Vector Machines, Nearest Neighbors, Gaussian Pro-
cesses, Decision Trees, Ensemble methods (Forests of randomised trees, Ad-
aBoost and Gradient Tree Boosting), Multi-layer Perceptrons (https://scikit-
learn.org/stable/).
Python CNTK A framework, developed by Microsoft, that provides tools for building ML
and DL models (https://docs.microsoft.com/en-us/cognitive-toolkit/).
Python PyTorch A framework, developed by Facebook, for building ML and DL models
(https://pytorch.org/).
2.6.11. Clustering-based forecasting R tsfknn tsfknn Time Series Forecasting Using Nearest Neighbours.
2.6.12. Hybrid methods Python ESRNN-GPU A GPU-enabled version of the hybrid model used by the winner of M4 com-
petition (https://github.com/damitkwr/ESRNN-GPU).
2.7.1. Parametric methods for intermittent demand R tsintermittent crost; tsb Parametric forecasting methods for intermitternt demand.
forecasting
R forecast croston Forecasts for intermittent demand using Croston’s method.
2.7.2. Non-parametric intermittent demand meth- R tsintermittent imapa MAPA for intermittent demand data.
ods
2.7.3. Classification methods R tsintermittent idclass Time series categorisation for intermittent demand.
R tsutils abc; xyz; abcxyz Classification functions and routines.
2.8.3 Forecasting with text information Python NLTK The Natural Language Toolkit in Python
Python SpaCy An open source library for advanced Natural Language Processing in
Python.
2.9.1. Cross-sectional hierarchical forecasting R hts Functions and routines for hierarchical and grouped time series forecasting.
 Related section Software Package/Library/Toolbox Function(s) Comments
2.9.2. Temporal aggregation R MAPA mapa; mapasimple Functions and wrappers for using the Multiple Aggregation Prediction Al-
gorithm (MAPA) for time series forecasting.
R thief thief Temporal Hierarchical Forecasting.
R tsintermittent imapa MAPA for intermittent demand data with automatic model selection based
on the PK classification.
2.9.4. Ecological inference forecasting R eiPack ei.MD.bayes; ei.reg; Returns local and global forecasts of inner cells in R×C tables under a Multi-
ei.reg.bayes nomial Dirichlet model or using ecological regression.
R lphom lphom Returns global forecasts of inner cells of a RxC table using the lphom model
(https://go.uv.es/N6Wnx75).
2.11.1. Point, interval and pHDR forecast error R forecast accuracy Accuracy measures for a forecast model.
measures
2.11.2. Evaluating probabilistic forecasts R scoringRules Scoring rules for parametric and simulated distribution forecasts.
R verification crps Continuous ranked probability score.
2.11.4. Statistical tests of forecast performance R forecast dm.test Diebold-Mariano test for predictive accuracy.
R tsutils nemenyi Nonparametric multiple comparisons (Nemenyi test).
171
 Appendix C. Data sets

Table A.2: A list of indicative publicly available data sets.

Related section Description Link

2.6.7. Deep Probabilistic Models Data for wikipedia page views, Dominicks retail, elec- https://gluon-ts.mxnet.io/api/gluonts/gluonts.dataset.repository.datasets.html
tricity consumption, traffic lane occupation
2.8.3. Forecasting with text Information Movie reviews data provided by the Stanford NLP https://nlp.stanford.edu/sentiment/code.html
group
3.2.3. Forecasting for inventories Grupo Bimbo Inventory Demand https://www.kaggle.com/c/grupo-bimbo-inventory-demand
3.2.4. Retail sales forecasting Rossmann Store Sales https://www.kaggle.com/c/rossmann-store-sales
Corporación Favorita Grocery Sales Forecasting https://www.kaggle.com/c/favorita-grocery-sales-forecasting
Walmart Recruiting – Store Sales Forecasting https://www.kaggle.com/c/walmart-recruiting-store-sales-forecasting
Walmart Recruiting II: Sales in Stormy Weather https://www.kaggle.com/c/walmart-recruiting-sales-in-stormy-weather
Store Item Demand Forecasting Challenge https://www.kaggle.com/c/demand-forecasting-kernels-only
3.2.5. Online retail forecasting Online Product Sales https://www.kaggle.com/c/online-sales
3.2.9. Predictive maintenance Robot Execution Failures https://archive.ics.uci.edu/ml/datasets/Robot+Execution+Failures
Gearbox Fault Detection https://c3.nasa.gov/dashlink/resources/997/
Air Pressure System Failure at Scania Trucks https://archive.ics.uci.edu/ml/datasets/IDA2016Challenge
Generic, Scalable and Decentralised Fault Detection https://zenodo.org/record/831471##.WwQIPUgvxPY
for Robot Swarms
172

Wind turbine data (e.g., failures) https://opendata.edp.com/pages/homepage/

3.3. Economics and finance Two Sigma Financial Modelling Challenge https://www.kaggle.com/c/two-sigma-financial-modeling/overview
Financial, economic, and alternative data sets, serving https://www.quandl.com/
investment professionals.
3.3.2. Forecasting GDP and Inflation Repository website with Dynare codes and data sets to https://github.com/johannespfeifer/dsge mod
estimate different DSGE models and use them for fore-
casting
Data set for Macroeconomic variables for US economy https://fred.stlouisfed.org/
Data set for Macroeconomic variables for OECD econ- https://data.oecd.org/
omy
Data set for Macroeconomic variables for EU economy https://ec.europa.eu/eurostat/data/database
3.3.8. House price forecasting Zillow Prize: Zillow’s Home Value Prediction (Zesti- https://www.kaggle.com/c/zillow-prize-1
mate)
Sberbank Russian Housing Market https://www.kaggle.com/c/sberbank-russian-housing-market
Western Australia Rental Prices https://www.kaggle.com/c/deloitte-western-australia-rental-prices
3.3.13. Forecasting returns to investment style Algorithmic Trading Challenge https://www.kaggle.com/c/AlgorithmicTradingChallenge
3.3.14. Forecasting stock returns The Winton Stock Market Challenge https://www.kaggle.com/c/the-winton-stock-market-challenge
The Big Data Combine Engineered by BattleFin https://www.kaggle.com/c/battlefin-s-big-data-combine-forecasting-challenge/data
3.4. Energy VSB Power Line Fault Detection https://www.kaggle.com/c/vsb-power-line-fault-detection
ASHRAE – Great Energy Predictor III https://www.kaggle.com/c/ashrae-energy-prediction
3.4.4. Hybrid machine learning system for short-term Global Energy Forecasting Competition 2012 – Load https://www.kaggle.com/c/global-energy-forecasting-competition-2012-load-forecasting
load forecasting Forecasting
3.4.8. Wind power forecasting Global Energy Forecasting Competition 2012 – Wind https://www.kaggle.com/c/GEF2012-wind-forecasting
Forecasting
 Related section Description Link
3.4.9. Forecasting with simulation in energy planning Brazilian National Electric Systems Operator (hydro, http://www.ons.org.br/Paginas/resultados-da-operacao/historico-da-
solar, wind, nuclear and thermal generation data) operacao/geracao energia.aspx
3.4.10. Power forecasting from PV systems Power measurements from a PV power plant and grid https://doi.org/10.25747/edf8-m258
of numerical weather predictions
AMS 2013-2014 Solar Energy Prediction Contest https://www.kaggle.com/c/ams-2014-solar-energy-prediction-contest
SolarTechLab data set https://ieee-dataport.org/open-access/photovoltaic-power-and-weather-parameters
3.4.12. Collaborative forecasting in the energy sector Solar power time series from 44 small-scale PV in https://doi.org/10.25747/gywm-9457
Évora, Portugal
Wind power time series from different owners (geo- https://doi.org/10.15129/9e1d9b96-baa7-4f05-93bd-99c5ae50b141
graphically distributed)
Australian Electricity Market Operator (AEMO) 5 https://data.london.gov.uk/dataset/smartmeter-energy-use-data-in-london-households
Minute Wind Power Data
Electric vehicles charging data (arrivals, departures, https://eatechnology.com/consultancy-insights/my-electric-avenue/
current, voltage, etc.)
3.5.2. Weather forecasting How Much Did It Rain? https://www.kaggle.com/c/how-much-did-it-rain-ii
3.5.3. Air quality forecasting EMC Data Science Global Hackathon (Air Quality Pre- https://www.kaggle.com/c/dsg-hackathon/overview
diction)
3.6. Social good and demographic forecasting LANL Earthquake Prediction https://www.kaggle.com/c/LANL-Earthquake-Prediction
3.6.1. Healthcare Flu Forecasting https://www.kaggle.com/c/genentech-flu-forecasting
West Nile Virus Prediction https://www.kaggle.com/c/predict-west-nile-virus
3.6.2. Epidemics and pandemics COVID-19 Data Repository by the Center for Systems https://github.com/CSSEGISandData/COVID-19
173

Science and Engineering (CSSE) at Johns Hopkins Uni-

versity
3.6.4. Forecasting mortality data Human Mortality Database https://www.mortality.org
EuroMOMO https://www.euromomo.eu/
The Economist https://github.com/TheEconomist/covid-19-excess-deaths-tracker
The New York Times https://github.com/Financial-Times/coronavirus-excess-mortality-data
The Financial Times https://github.com/nytimes/covid-19-data/tree/master/excess-deaths
ANACONDA- Quality assessment of mortality data https://crvsgateway.info/anaconda
3.6.5. Annuity price forecasting Human Mortality Database https://www.mortality.org
Australian Human Mortality Database https://demography.cass.anu.edu.au/research/australian-human-mortality-database
Canadian Human Mortality Database http://www.bdlc.umontreal.ca/CHMD/
French Human Mortality Database https://frdata.org/fr/french-human-mortality-database
United States Mortality Database https://usa.mortality.org
3.8. Other applications Forecast Eurovision Voting https://www.kaggle.com/c/Eurovision2010
Reducing Commercial Aviation Fatalities https://www.kaggle.com/c/reducing-commercial-aviation-fatalities
Porto Seguro’s Safe Driver Prediction https://www.kaggle.com/c/porto-seguro-safe-driver-prediction
Recruit Restaurant Visitor Forecasting https://www.kaggle.com/c/recruit-restaurant-visitor-forecasting
Restaurant Revenue Prediction https://www.kaggle.com/c/restaurant-revenue-prediction
Coupon Purchase Prediction https://www.kaggle.com/c/coupon-purchase-prediction
Bike Sharing Demand https://www.kaggle.com/c/bike-sharing-demand
Google Analytics Customer Revenue Prediction https://www.kaggle.com/c/ga-customer-revenue-prediction
Santander Value Prediction Challenge https://www.kaggle.com/c/santander-value-prediction-challenge
Santander Customer Transaction Prediction https://www.kaggle.com/c/santander-customer-transaction-prediction
 Related section Description Link
3.8. Other applications (continued) Acquire Valued Shoppers Challenge https://www.kaggle.com/c/acquire-valued-shoppers-challenge
Risky Business https://www.kaggle.com/c/risky-business
Web Traffic Time Series Forecasting https://www.kaggle.com/c/web-traffic-time-series-forecasting
A repository of data sets, including time series ones, https://github.com/awesomedata/awesome-public-datasets
that can be used for benchmarking forecasting meth-
ods in various applications of interest.
WSDM – KKBox’s Churn Prediction Challenge https://www.kaggle.com/c/kkbox-churn-prediction-challenge/overview
Homesite Quote Conversion https://www.kaggle.com/c/homesite-quote-conversion
Liberty Mutual Group: Property Inspection Prediction https://www.kaggle.com/c/liberty-mutual-group-property-inspection-prediction
Liberty Mutual Group – Fire Peril Loss Cost https://www.kaggle.com/c/liberty-mutual-fire-peril
A set of more than 490,000 time series (micro, macro, http://fsudataset.com/
demographic, finance, other) to download.
A repository of data sets, including time series ones, https://github.com/awesomedata/awesome-public-datasets
that can be used for benchmarking forecasting meth-
ods in various applications of interest.
3.8.1. Tourism demand forecasting TourMIS: Database with annual and monthly tourism https://www.tourmis.info/index e.html
time series (e.g., arrivals and bednights) covering Eu-
ropean countries, regions, and cities (free registration
required)
Tourism Forecasting https://www.kaggle.com/c/tourism1; https://www.kaggle.com/c/tourism2
3.8.2. Forecasting for airports Airline and Airport performance data provided by the https://www.transtats.bts.gov/
174

U.S. Department of Transportation

3.8.3. Traffic flow forecasting New York City Taxi Fare Prediction https://www.kaggle.com/c/demand-forecasting-kernels-only
RTA Freeway Travel Time Prediction https://www.kaggle.com/c/RTA
ECML/PKDD 15: Taxi Trip Time Prediction https://www.kaggle.com/c/pkdd-15-taxi-trip-time-prediction-ii
BigQuery-Geotab Intersection Congestion https://www.kaggle.com/c/bigquery-geotab-intersection-congestion/overview
Traffic volume counts collected by DOT for New York https://data.cityofnewyork.us/Transportation/Traffic-Volume-Counts-2014-2019-/ertz-
Metropolitan Transportation Council hr4r
3.8.5. Elections forecasting Party registration in South-East North Carolina (senc https://www2.ncleg.net/RnR/Redistricting/BaseData2001
data, eiPack R package)
2002 New Zealand General Election – Official results https://www.electionresults.govt.nz/electionresults 2002/e9/html/statistics.html
and statistics
3.8.6. Sports forecasting NFL Big Data Bowl https://www.kaggle.com/c/nfl-big-data-bowl-2020
3.8.9. Safety and security Microsoft Malware Prediction https://www.kaggle.com/c/microsoft-malware-prediction
3.8.13. Forecasting competitions Data for the M, M2, M3 and M4 forecasting competi- https://forecasters.org/resources/time-series-data/
tions
Time Series Competition Data (R package) https://github.com/robjhyndman/tscompdata
Mcomp: Data for the M and M3 forecasting competi- https://cran.r-project.org/package=Mcomp
tions (R package)
M4comp2018: Data for the M4 forecasting competi- https://github.com/carlanetto/M4comp2018
tion (R package)
Data for the M4 forecasting competition (csv files) https://github.com/Mcompetitions/M4-methods/tree/master/Dataset
Tcomp: Data from the 2010 Tourism forecasting com- https://cran.r-project.org/package=Tcomp
petition (R package)
 Related section Description Link
175 3.8.13. Forecasting competitions (continued) Data for the M5 forecasting competition (csv files) https://github.com/Mcompetitions/M5-methods/tree/master/Dataset
References
Aaltonen, K., 2011. Project stakeholder analysis as an environmental interpretation process. International Journal of Project Manage-
ment 29 (2), 165–183.
Aastveit, K. A., Mitchell, J., Ravazzolo, F., van Dijk, H. K., 2019. The evolution of forecast density combinations in economics.
Abadi, M., Chu, A., Goodfellow, I., McMahan, H. B., Mironov, I., Talwar, K., Zhang, L., 2016. Deep learning with differential privacy.
In: Proc. of the 2016 ACM SIGSAC Conference on Computer and Communications Security (ACM CCS). pp. 308–318.
Abraham, B., Chuang, A., 1989. Outlier detection and time series modeling. Technometrics 31 (2), 241–248.
Abraham, B., George E. P. Box, 1979. Bayesian analysis of some outlier problems in time series. Biometrika 66 (2), 229–236.
Abramson, B., Finizza, A., 1991. Using belief networks to forecast oil prices. International Journal of Forecasting 7 (3), 299–315.
Abramson, B., Finizza, A., 1995. Probabilistic forecasts from probabilistic models: a case study in the oil market. International Journal
of forecasting 11 (1), 63–72.
Abramson, G., Zanette, D., 1998. Statistics of extinction and survival in Lotka–Volterra systems. Physical Review E 57, 4572–4577.
Acemoglu, D., Scott, A., 1994. Asymmetries in the cyclical behaviour of UK labour markets. The Economic Journal 104 (427), 1303–
1323.
Achen, C. H., Phillips Shively, W., 1995. Cross-Level Inference. University of Chicago Press.
Acquaviva, A., Apiletti, D., Attanasio, A., Baralis, E., Castagnetti, F. B., Cerquitelli, T., Chiusano, S., Macii, E., Martellacci, D., Patti, E.,
2015. Enhancing energy awareness through the analysis of thermal energy consumption. In: EDBT/ICDT Workshops. pp. 64–71.
Afanasyev, D. O., Fedorova, E. A., 2019. On the impact of outlier filtering on the electricity price forecasting accuracy. Applied Energy
236, 196–210.
Agarwal, A., Dahleh, M., Sarkar, T., 2019. A marketplace for data: An algorithmic solution. In: Proceedings of the 2019 ACM
Conference on Economics and Computation. pp. 701–726.
Aggarwal, C., Zhai, C., 2012. Mining text data. Springer Science & Business Media.
Agnolucci, P., 2009. Volatility in crude oil futures: A comparison of the predictive ability of GARCH and implied volatility models.
Energy Economics 31 (2), 316–321.
Ahlburg, D. A., Vaupel, J. W., 1990. Alternative projections of the U.S. population. Demography 27 (4), 639–652.
Ahmad, A., Hassan, M., Abdullah, M., Rahman, H., Hussin, F., Abdullah, H., Saidur, R., 2014. A review on applications of ANN and
SVM for building electrical energy consumption forecasting. Renewable and Sustainable Energy Reviews 33, 102–109.
Ahmad, A., Javaid, N., Mateen, A., Awais, M., Khan, Z. A., 2019. Short-Term load forecasting in smart grids: An intelligent modular
approach. Energies 12 (1), 164.
Ahmad, M. W., Mourshed, M., Rezgui, Y., 2017. Trees vs neurons: Comparison between random forest and ANN for high-resolution
prediction of building energy consumption. Energy and Buildings 147, 77–89.
Ahmed, R., Sreeram, V., Mishra, Y., Arif, 2020. A review and evaluation of the state-of-the-art in PV solar power forecasting: Tech-
niques and optimization. Renewable and Sustainable Energy Reviews 124, 109792.
Aı̈t-Sahalia, Y., Cacho-Diaz, J., Laeven, R. J. A., 2015. Modeling financial contagion using mutually exciting jump processes. Journal
of Financial Economics 117 (3), 585–606.
Aizenman, J., Jinjarak, Y., 2013. Real Estate Valuation, Current Account and Credit Growth Patterns, Before and After the 2008-9
Crisis. Working Paper 19190, National Bureau of Economic Research.
Akaike, H., 1973. Information theory and an extension of the maximum likelihood principle. In: Petrov, B. N., Csáki, F. (Eds.),
Proceedings of the Second International Symposium on Information Theory. Csáki, Budapest, pp. 267–281.
Akram, F., Binning, A., Maih, J., 2015. Joint prediction bands for macroeconomic risk management. Tech. Rep. No 5/2016, Centre for
Applied Macro- and Petroleum economics (CAMP) Working Paper Series.
Aksin, Z., Armony, M., Mehrotra, V., 2007. The modern call center: A multi-disciplinary perspective on operations management
research. Production and Operations Management 16 (6), 665–688.
Aktekin, T., Soyer, R., 2011. Call center arrival modeling: A Bayesian state-space approach. Naval Research Logistics 58 (1), 28–42.
Al-Azzani, M. A., Davari, S., England, T. J., 2020. An empirical investigation of forecasting methods for ambulance calls-a case study.
Health Systems, 1–18.
Al-Homoud, M. S., 2001. Computer-aided building energy analysis techniques. Building and Environment 36 (4), 421–433.
Albon, C., 2018. Python Machine Learning Cookbook. O’Reilly UK Ltd.
Albulescu, C. T., Tiwari, A. K., Ji, Q., 2020. Copula-based local dependence among energy, agriculture and metal commodities
markets. Energy 202, 117762.
Aldor-Noiman, S., Feigin, P. D., Mandelbaum, A., 2009. Workload forecasting for a call center: Methodology and a case study. The
Annals of Applied Statistics 3 (4), 1403–1447.

176
Alexandrov, A., Benidis, K., Bohlke-Schneider, M., Flunkert, V., Gasthaus, J., Januschowski, T., Maddix, D. C., Rangapuram, S.,
Salinas, D., Schulz, J., Stella, L., Türkmen, A. C., Wang, Y., 2019. GluonTS: Probabilistic time series models in python. Journal of
Machine Learning Research.
Alho, J. M., Spencer, B. D., 2005. Statistical Demography and Forecasting. Springer, New York.
Ali, M. M., Boylan, J. E., 2011. Feasibility principles for downstream demand inference in supply chains. Journal of the Operational
Research Society 62 (3), 474–482.
Ali, M. M., Boylan, J. E., Syntetos, A. A., 2012. Forecast errors and inventory performance under forecast information sharing.
International Journal of Forecasting 28 (4), 830–841.
Alizadeh, S., Brandt, M. W., Diebold, F. X., 2002. Range-based estimation of stochastic volatility models. Journal of Finance 57 (3),
1047–1091.
Almeida, C., Czado, C., 2012. Efficient Bayesian inference for stochastic time-varying copula models. Computational Statistics & Data
Analysis 56 (6), 1511–1527.
Aloui, R., Hammoudeh, S., Nguyen, D. K., 2013. A time-varying copula approach to oil and stock market dependence: The case of
transition economies. Energy Economics 39, 208–221.
Alquist, R., Bhattarai, S., Coibion, O., 2020. Commodity-price comovement and global economic activity. Journal of Monetary Eco-
nomics 112, 41–56.
Alquist, R., Kilian, L., Vigfusson, R. J., 2013. Forecasting the price of oil. In: Handbook of Economic Forecasting. Vol. 2. Elsevier, pp.
427–507.
Alvarado-Valencia, J., Barrero, L. H., Önkal, D., Dennerlein, J. T., 2017. Expertise, credibility of system forecasts and integration
methods in judgmental demand forecasting. International Journal of Forecasting 33 (1), 298–313.
Alvarado-Valencia, J. A., Barrero, L. H., 2014. Reliance, trust and heuristics in judgmental forecasting. Computers in Human Behavior
36, 102–113.
Alvarez-Ramirez, J., Soriano, A., Cisneros, M., Suarez, R., 2003. Symmetry/anti-symmetry phase transitions in crude oil markets.
Physica A: Statistical Mechanics and its Applications 322, 583–596.
Amarasinghe, A., Wichmann, O., Margolis, H. S., Mahoney, R. T., 2010. Forecasting dengue vaccine demand in disease endemic and
non-endemic countries. Human Vaccines 6 (9), 745–753.
Amendola, A., Niglio, M., Vitale, C., 2006. The moments of SETARMA models. Statistics & Probability Letters 76 (6), 625–633.
Amisano, G., Giacomini, R., 2007. Comparing density forecasts via weighted likelihood ratio tests. Journal of Business & Economic
Statistics 25 (2), 177–190.
An, S., Schorfheide, F., 2007. Bayesian analysis of DSGE models. Econometric Reviews 26 (2-4), 113–172.
Anderson, B. D. O., Moore, J. B., 1979. Optimal Filtering. Prentice-Hall: Englewood Cliffs, NJ.
Anderson, J. L., 1996. A method for producing and evaluating probabilistic forecasts from ensemble model integrations. Journal of
Climate 9, 1518–1530.
Anderson, V. O., Nochmals, U., 1914. The elimination of spurious correlation due to position in time or space. Biometrika 10 (2/3),
269–279.
Andersson, E., Kühlmann-Berenzon, S., Linde, A., Schiöler, L., Rubinova, S., Frisén, M., 2008. Predictions by early indicators of the
time and height of the peaks of yearly influenza outbreaks in sweden. Scandinavian Journal of Public Health 36 (5), 475–482.
Andrade, J., Filipe, J., Reis, M., Bessa, R., 2017. Probabilistic price forecasting for day-ahead and intraday markets: Beyond the
statistical model. Sustainability 9 (1990), 1–29.
Andrawis, R. R., Atiya, A. F., El-Shishiny, H., 2011. Combination of long term and short term forecasts, with application to tourism
demand forecasting. International Journal of Forecasting 27 (3), 870–886.
Andrés, M. A., Peña, D., Romo, J., 2002. Forecasting time series with sieve bootstrap. Journal of Statistical Planning and Inference
100 (1), 1–11.
Andrews, B. H., Cunningham, S. M., 1995. Ll bean improves call-center forecasting. Interfaces 25 (6), 1–13.
Andrews, R. L., Currim, I. S., Leeflang, P., Lim, J., 2008. Estimating the SCAN* PRO model of store sales: HB, FM or just OLS?
International Journal of Research in Marketing 25 (1), 22–33.
Aneiros-Pérez, G., Vieu, P., 2008. Nonparametric time series prediction: A semi-functional partial linear modeling. Journal of Multi-
variate Analysis 99 (5), 834–857.
Ang, A., Bekaert, G., 2002a. Regime switches in interest rates. Journal of Business & Economic Statistics 20 (2), 163–182.
Ang, A., Bekaert, G., 2002b. Short rate nonlinearities and regime switches. Journal of Economic Dynamics & Control 26 (7), 1243–
1274.
Ang, A., Bekaert, G., Wei, M., 2008. The term structure of real rates and expected inflation. The Journal of Finance 63 (2), 797–849.

177
Angelini, G., De Angelis, L., 2019. Efficiency of online football betting markets. International Journal of Forecasting 35 (2), 712–721.
Angus, J. E., 1992. Asymptotic theory for bootstrapping the extremes. Communications in Statistics-Theory and Methods 22 (1),
15–30.
Anselin, L., Tam Cho, W. K., 2002. Spatial effects and ecological inference. Political Analysis 10 (3), 276–297.
Antipov, A., Meade, N., 2002. Forecasting call frequency at a financial services call centre. Journal of the Operational Research Society
53 (9), 953–960.
Antonakakis, N., Chatziantoniou, I., Floros, C., Gabauer, D., 2018. The dynamic connectedness of U.K. regional property returns.
Urban Studies 55 (14), 3110–3134.
Apiletti, D., Baralis, E., Cerquitelli, T., Garza, P., Michiardi, P., Pulvirenti, F., 2015. Pampa-hd: A parallel mapreduce-based frequent
pattern miner for high-dimensional data. In: 2015 IEEE International Conference on Data Mining Workshop (ICDMW). IEEE, pp.
839–846.
Apiletti, D., Baralis, E., Cerquitelli, T., Garza, P., Pulvirenti, F., Michiardi, P., 2017. A parallel mapreduce algorithm to efficiently
support itemset mining on high dimensional data. Big Data Research 10, 53–69.
Apiletti, D., Pastor, E., 2020. Correlating espresso quality with coffee-machine parameters by means of association rule mining.
Electronics 9 (1), 100.
Apiletti, D., Pastor, E., Callà, R., Baralis, E., 2020. Evaluating espresso coffee quality by means of time-series feature engineering. In:
EDBT/ICDT Workshops.
Archak, N., Ghose, A., Ipeirotis, P., 2011. Deriving the pricing power of product features by mining consumer reviews. Management
Science 57 (8), 1485–1509.
Archer, C., Baptista, A., Leen, T. K., 2003. Fault detection for salinity sensors in the columbia estuary. Water Resources Research
39 (3).
Arinze, B., 1994. Selecting appropriate forecasting models using rule induction. Omega 22 (6), 647–658.
Arlot, S., Celisse, A., 2010. A survey of cross-validation procedures for model selection. Statistics Surveys 4, 40–79.
Armstrong, C., 2017. Omnichannel retailing and demand planning. The Journal of Business Forecasting 35 (4), 10–15.
Armstrong, J. S., 2001a. Combining forecasts. In: Principles of Forecasting. International Series in Operations Research & Manage-
ment Science. Springer, Boston, MA, pp. 417–439.
Armstrong, J. S., 2001b. Principles of forecasting: a handbook for researchers and practitioners. Springer Science & Business Media.
Armstrong, J. S., Green, K. C., 2018. Forecasting methods and principles: Evidence-based checklists. Journal of Global Scholars of
Marketing Science 28 (2), 103–159.
Armstrong, J. S., Green, K. C., Graefe, A., 2015. Golden rule of forecasting: Be conservative. Journal of Business Research 68 (8),
1717–1731.
Arnott, R. D., Beck, N., Kalesnik, V., West, J., 2016. How can ‘smart beta’ go horribly wrong? SSRN:3040949.
Aron, J., Muellbauer, J., 2020. Measuring excess mortality: the case of England during the Covid-19 pandemic.
https://www.oxfordmartin.ox.ac.uk/publications/measuring-excess-mortality-the-case-of-england-during-the-covid-19-
pandemic/, accessed on 2020-08-20.
Arora, S., Taylor, J. W., Mak, H.-Y., 2020. Probabilistic forecasting of patient waiting times in an emergency department.
arXiv:2006.00335.
Arrhenius, S. A., 1896. On the influence of carbonic acid in the air upon the temperature of the ground. London, Edinburgh, and
Dublin Philosophical Magazine and Journal of Science (fifth series) 41, 237–275.
Artis, M., Marcellino, M., 2001. Fiscal forecasting: The track record of the imf, oecd and ec. The Econometrics Journal 4 (1), S20–S36.
Arunraj, N. S., Ahrens, D., 2015. A hybrid seasonal autoregressive integrated moving average and quantile regression for daily food
sales forecasting. International Journal of Production Economics 170, 321–335.
Arvan, M., Fahimnia, B., Reisi, M., Siemsen, E., 2019. Integrating human judgement into quantitative forecasting methods: A review.
Omega 86, 237–252.
Asai, M., 2013. Heterogeneous asymmetric dynamic conditional correlation model with stock return and range. Journal of Forecasting
32 (5), 469–480.
Asai, M., Brugal, I., 2013. Forecasting volatility via stock return, range, trading volume and spillover effects: The case of Brazil. North
American Journal of Economics and Finance 25, 202–213.
Asimakopoulos, S., Dix, A., 2013. Forecasting support systems technologies-in-practice: A model of adoption and use for product
forecasting. International Journal of Forecasting 29 (2), 322–336.
Asimakopoulos, S., Paredes, J., Warmedinger, T., 2020. Real-time fiscal forecasting using mixed-frequency data. The Scandinavian
Journal of Economics 122, 369–390.

178
Askanazi, R., Diebold, F. X., Schorfheide, F., Shin, M., 2018. On the comparison of interval forecasts. Journal of Time Series Analysis
39 (6), 953–965.
Asness, C. S., 2016. Invited editorial comment: The siren song of factor timing aka “smart beta timing” aka “style timing”. Journal of
Portfolio Management 42 (5), 1–6.
Assimakopoulos, V., Nikolopoulos, K., 2000. The Theta model: a decomposition approach to forecasting. International Journal of
Forecasting 16 (4), 521–530.
Athanasopoulos, G., Ahmed, R. A., Hyndman, R. J., 2009. Hierarchical forecasts for Australian domestic tourism. International
Journal of Forecasting 25 (1), 146–166.
Athanasopoulos, G., Hyndman, R. J., Kourentzes, N., Petropoulos, F., 2017. Forecasting with temporal hierarchies. European Journal
of Operational Research 262 (1), 60–74.
Athanasopoulos, G., Hyndman, R. J., Song, H., Wu, D. C., 2011. The tourism forecasting competition. International Journal of Fore-
casting 27 (3), 822–844.
Athanasopoulos, G., Song, H., Sun, J. A., 2018. Bagging in tourism demand modeling and forecasting. Journal of Travel Research
57 (1), 52–68.
Athey, S., 2018. The impact of machine learning on economics. In: Agrawal, A., Gans, J., Goldfarb, A. (Eds.), The Economics of
Artificial Intelligence: An Agenda. University of Chicago Press, pp. 507–547.
Aue, A., Norinho, D. D., Hörmann, S., 2015. On the prediction of stationary functional time series. Journal of the American Statistical
Association 110 (509), 378–392.
Austin, C., Kusumoto, F., 2016. The application of big data in medicine: current implications and future directions. Journal of
Interventional Cardiac Electrophysiology 47 (1), 51–59.
Avramidis, A. N., Deslauriers, A., L’Ecuyer, P., 2004. Modeling daily arrivals to a telephone call center. Management Science 50 (7),
896–908.
Axelrod, R., 1997. Advancing the art of simulation in the social sciences. In: Simulating Social Phenomena. Springer, pp. 21–40.
Ayton, P., Önkal, D., McReynolds, L., 2011. Effects of ignorance and information on judgments and decisions. Judgment and Decision
Making 6 (5), 381–391.
Baade, R. A., Matheson, V. A., 2016. Going for the Gold: The economics of the Olympics. Journal of Economic Perspectives 30 (2),
201–18.
Baardman, L., Levin, I., Perakis, G., Singhvi, D., 2018. Leveraging comparables for new product sales forecasting. Production and
Operations Management 27 (12), 2340–2343.
Babai, M. Z., Ali, M. M., Nikolopoulos, K., 2012. Impact of temporal aggregation on stock control performance of intermittent demand
estimators: Empirical analysis. Omega 40 (6), 713–721.
Babai, M. Z., Dallery, Y., Boubaker, S., Kalai, R., 2019. A new method to forecast intermittent demand in the presence of inventory
obsolescence. International Journal of Production Economics 209, 30–41.
Babai, M. Z., Syntetos, A., Teunter, R., 2014. Intermittent demand forecasting: An empirical study on accuracy and the risk of
obsolescence. International Journal of Production Economics 157, 212–219.
Babai, M. Z., Tsadiras, A., Papadopoulos, C., 2020. On the empirical performance of some new neural network methods for forecasting
intermittent demand. IMA Journal of Management Mathematics 31 (3), 281–305.
Babu, A., Levine, A., Ooi, Y. H., Pedersen, L. H., Stamelos, E., 2020. Trends everywhere. Journal of Investment Management 18 (1),
52–68.
Bacchetti, A., Saccani, N., 2012. Spare parts classification and demand forecasting for stock control: Investigating the gap between
research and practice. Omega 40 (6), 722–737.
Baccianella, S., Esuli, A., Sebastiani, F., 2010. Sentiwordnet 3.0: an enhanced lexical resource for sentiment analysis and opinion
mining. In: Proceedings of the Seventh International Conference on Language Resources and Evaluation (LREC’10). Vol. 10. pp.
2200–2204.
Bacha, H., Meyer, W., 1992. A neural network architecture for load forecasting. In: IJCNN International Joint Conference on Neural
Networks. Vol. 2. pp. 442–447.
Baecke, P., De Baets, S., Vanderheyden, K., 2017. Investigating the added value of integrating human judgement into statistical
demand forecasting systems. International Journal of Production Economics 191, 85–96.
Bai, J., Ng, S., 2008. Forecasting economic time series using targeted predictors. Journal of Econometrics 146(2), 304–317.
Bai, J., Ng, S., 2009. Boosting diffusion indices. Journal of Applied Econometrics 24, 607–629.
Baicker, K., Chandra, A., Skinner, J. S., 2012. Saving money or just saving lives? improving the productivity of US health care
spending. Annual Review of Economics 4 (1), 33–56.

179
Baillie, R. T., Bollerslev, T., 1992. Prediction in dynamic models with time-dependent conditional variances. Journal of Econometrics
1–2 (52), 91–113.
Baillie, R. T., Bollerslev, T., Mikkelsen, H. O., 1996. Fractionally integrated generalized autoregressive conditional heteroskedasticity.
Journal of Econometrics 74 (1), 3–30.
Baker, J., 2021. Maximizing forecast value added through machine learning and nudges. Foresight: The International Journal of
Applied Forecasting 60.
Baker, S. R., Bloom, N., Davis, S. J., 2016. Measuring economic policy uncertainty. The Quarterly Journal of Economics 131 (4),
1593–1636.
Balke, N. S., 1993. Detecting level shifts in time series. Journal of Business & Economic Statistics 11 (1), 81–92.
Balke, N. S., Fomby, T. B., 1997. Threshold cointegration. International Economic Review 38 (3), 627–645.
Bandara, K., Bergmeir, C., Hewamalage, H., 2020a. LSTM-MSNet: leveraging forecasts on sets of related time series with multiple
seasonal patterns. IEEE Transactions on Neural Networks and Learning Systems.
Bandara, K., Bergmeir, C., Smyl, S., 2020b. Forecasting across time series databases using recurrent neural networks on groups of
similar series: a clustering approach. Expert Systems with Applications 140, 112896.
Bandyopadhyay, S., 2009. A dynamic model of cross-category competition: theory, tests and applications. Journal of Retailing 85 (4),
468–479.
Bangwayo-Skeete, P. F., Skeete, R. W., 2015. Can Google data improve the forecasting performance of tourist arrivals? mixed-data
sampling approach. Tourism Management 46, 454–464.
Banks, J., Blundell, R., Oldfield, Z., Smith, J. P., 2015. House price volatility and the housing ladder. Working Paper 21255, National
Bureau of Economic Research.
Bannister, R. N., Chipilski, H. G., Martinez-Alvarado, O., 2020. Techniques and challenges in the assimilation of atmospheric water
observations for numerical weather prediction towards convective scales. Quarterly Journal of the Royal Meteorological Society
146 (726), 1–48.
Bansal, R., Tauchen, G., Zhou, H., 2004. Regime shifts, risk premiums in the term structure, and the business cycle. Journal of
Business & Economic Statistics 22 (4), 396–409.
Bansal, R., Zhou, H., 2002. Term structure of interest rates with regime shifts. The Journal of Finance 57 (5), 1997–2043.
Bao, Y., Lee, T.-H., Saltoglu, B., 2007. Comparing density forecast models. Journal of Forecasting 26 (3), 203–225.
Baptista, M., Sankararaman, S., de Medeiros, I. P., Nascimento, C., Prendinger, H., Henriques, E. M. P., 2018. Forecasting fault events
for predictive maintenance using data-driven techniques and ARMA modeling. Computers & Industrial Engineering 115, 41–53.
Barbetta, S., Coccia, G., Moramarco, T., Brocca, L., Todini, E., 2017. The multi temporal/multi-model approach to predictive uncer-
tainty assessment in real-time flood forecasting. Journal of Hydrology 551, 555–576.
BarclayHedge, 2018. Survey: Majority of hedge fund pros use AI/Machine learning in investment strategies. https://www.
barclayhedge.com/insider/barclayhedge-survey-majority-of-hedge-fund-pros-use-ai-machine-learning-in-investment-strategies,
accessed on 2020-09-01.
Barker, J., 2020. Machine learning in M4: What makes a good unstructured model? International Journal of Forecasting 36 (1),
150–155.
Barnhart, C., Cohn, A., 2004. Airline schedule planning: Accomplishments and opportunities. Manufacturing & Service Operations
Management 6 (1), 3–22.
Barnhart, C., Fearing, D., Vaze, V., 2014. Modeling passenger travel and delays in the national air transportation system. Operations
Research 62 (3), 580–601.
Barnichon, R., Garda, P., 2016. Forecasting unemployment across countries: The ins and outs. European Economic Review 84, 165–
183.
Barr, J., 2018. New – predictive scaling for EC2, powered by machine learning. https://aws.amazon.com/blogs/aws/
new-predictive-scaling-for-ec2-powered-by-machine-learning, accessed on 2020-09-01.
Barron, A. R., 1994. Approximation and estimation bounds for artificial neural networks. Machine Learning 14 (1), 115–133.
Barroso, P., 2015. Momentum has its moments. Journal of Financial Economics 116 (1), 111–121.
Barrow, D., Kourentzes, N., Sandberg, R., Niklewski, J., 2020. Automatic robust estimation for exponential smoothing: Perspectives
from statistics and machine learning. Expert Systems with Applications 160, 113637.
Barrow, D. K., Crone, S. F., 2016a. A comparison of adaboost algorithms for time series forecast combination. International Journal of
Forecasting 32 (4), 1103–1119.
Barrow, D. K., Crone, S. F., 2016b. Cross-validation aggregation for combining autoregressive neural network forecasts. International
Journal of Forecasting 32 (4), 1120–1137.

180
Barrow, D. K., Kourentzes, N., 2016. Distributions of forecasting errors of forecast combinations: implications for inventory manage-
ment. International Journal of Production Economics 177, 24–33.
Bartelsman, E. J., Kurz, C. J., Wolf, Z., 2011. Using Census Microdata to Forecast US Aggregate Productivity. Working paper.
Bartelsman, E. J., Wolf, Z., 2014. Forecasting aggregate productivity using information from firm-level data. Review of Economics
and Statistics 96 (4), 745–755.
Bartezzaghi, E., Verganti, R., Zotteri, G., 1999. A simulation framework for forecasting uncertain lumpy demand. International
Journal of Production Economics 59 (1), 499–510.
Bass, F., 1969. A new product growth model for consumer durables. Management Science 15, 215–227.
Bass, F., Krishnan, T., Jain, D., 1994. Why the Bass model fits without decision variables. Marketing Science 13, 203–223.
Bassamboo, A., Cui, R., Moreno, A., 2015. Wisdom of crowds in operations: Forecasting using prediction markets. SSRN:2679663.
Basu, S., Fernald, J. G., Oulton, N., Srinivasan, S., 2003. The case of the missing productivity growth, or does information technology
explain why productivity accelerated in the United States but not in the United Kingdom? NBER Macroeconomics Annual 18,
9–63.
Bates, J. M., Granger, C. W. J., 1969. The combination of forecasts. Journal of the Operational Research Society 20 (4), 451–468.
Baumeister, C., Guérin, P., Kilian, L., 2015. Do high-frequency financial data help forecast oil prices? The MIDAS touch at work.
International Journal of Forecasting 31 (2), 238–252.
Baumeister, C., Kilian, L., 2015. Forecasting the real price of oil in a changing world: a forecast combination approach. Journal of
Business & Economic Statistics 33 (3), 338–351.
Bańbura, M., Giannone, D., Reichlin, L., 2010. Large bayesian vector auto regressions. Journal of Applied Econometrics 25 (1), 71–92.
Beare, B. K., Seo, J., Seo, W., 2017. Cointegrated linear processes in Hilbert space. Journal of Time Series Analysis 38 (6), 1010–1027.
Becker, R., Hurn, S., Pavlov, V., 2008. Modelling spikes in electricity prices. The Economic Record 83 (263), 371–382.
Beckmann, J., Schussler, R., 2016. Forecasting exchange rates under parameter and model uncertainty. Journal of International Money
and Finance 60, 267–288.
Behera, M. K., Majumder, I., Nayak, N., 2018. Solar photovoltaic power forecasting using optimized modified extreme learning
machine technique. Engineering Science and Technology an International Journal 21 (3).
Bekaert, G., Hodrick, R. J., Marshall, D. A., 2001. Peso problem explanations for term structure anomalies. Journal of Monetary
Economics 48 (2), 241–270.
Bekiros, S., Cardani, R., Paccagnini, A., Villa, S., 2016. Dealing with financial instability under a DSGE modeling approach with
banking intermediation: A predictability analysis versus TVP-VARs. Journal of Financial Stability 26 (C), 216–227.
Bekiros, S., Paccagnini, A., 2015a. Estimating point and density forecasts for the US economy with a factor-augmented vector autore-
gressive DSGE model. Studies in Nonlinear Dynamics & Econometrics 19 (2), 107–136.
Bekiros, S. D., Paccagnini, A., 2014. Bayesian forecasting with small and medium scale factor-augmented vector autoregressive DSGE
models. Computational Statistics & Data Analysis 71 (C), 298–323.
Bekiros, S. D., Paccagnini, A., 2015b. Macroprudential Policy And Forecasting Using Hybrid DSGE Models With Financial Frictions
And State Space Markov-Switching Tvp-Vars. Macroeconomic Dynamics 19 (7), 1565–1592.
Bekiros, S. D., Paccagnini, A., 2016. Policy Oriented Macroeconomic Forecasting with Hybrid DGSE and Time Varying Parameter
VAR Models. Journal of Forecasting 35 (7), 613–632.
Beliën, J., Forcé, H., 2012. Supply chain management of blood products: A literature review. European Journal of Operational Re-
search 217 (1), 1–16.
Bell, F., Smyl, S., 2018. Forecasting at Uber: An introduction. Accessed on 2020-09-02.
URL https://eng.uber.com/forecasting-introduction/
Ben-Gal, I., 2005. Outlier detection. In: Data mining and knowledge discovery handbook. Springer, pp. 131–146.
Ben Taieb, S., Taylor, J. W., Hyndman, R. J., 2020. Hierarchical probabilistic forecasting of electricity demand with smart meter data.
Journal of the American Statistical Association.
Benati, L., 2007. Drift and breaks in labor productivity. Journal of Economic Dynamics and Control 31 (8), 2847–2877.
Bender, J., Sun, X., Thomas, R., Zdorovtsov, V., 2018. The promises and pitfalls of factor timing. Journal of Portfolio Management
44 (4), 79–92.
Bendre, M., Manthalkar, R., 2019. Time series decomposition and predictive analytics using MapReduce framework. Expert Systems
with Applications 116, 108–120.
Bennell, J., Sutcliffe, C., 2004. Black-Scholes versus artificial neural networks in pricing FTSE 100 options. Intelligent Systems in
Accounting, Finance & Management 12 (4), 243–260.
Berdugo, V., Chaussin, C., Dubus, L., Hebrail, G., Leboucher, V., 2011. Analog method for collaborative very-short-term forecasting

181
 of powergeneration from photovoltaic systems. In: Next Generation Data Mining Summit. Athens, Greece, pp. 1–5.
Berg, J. E., Nelson, F. D., Rietz, T. A., 2008. Prediction market accuracy in the long run. International Journal of Forecasting 24 (2),
285–300.
Berger, J. O., 1985. Statistical Decision Theory and Bayesian Analysis. Springer.
Bergmeir, C., Benı́tez, J. M., 2012. On the use of cross-validation for time series predictor evaluation. Information Sciences 191,
192–213.
Bergmeir, C., Hyndman, R. J., Benı́tez, J. M., 2016. Bagging exponential smoothing methods using STL decomposition and Box–Cox
transformation. International Journal of Forecasting 32 (2), 303–312.
Bergmeir, C., Hyndman, R. J., Koo, B., 2018. A note on the validity of cross-validation for evaluating autoregressive time series
prediction. Computational Statistics & Data Analysis 120, 70–83.
Berkowitz, J., 2001. Testing density forecasts, with applications to risk management. Journal of Business & Economic Statistics 19 (4),
465–474.
Berlinski, D., 2009. The Devil’s Delusion: Atheism and its Scientific Pretensions. Basic Books.
Bernard, A., Busse, M., 2004. Who wins the Olympic Games: Economic resources and medal totals. Review of Economics and Statistics
86 (1), 413–417.
Bernardini Papalia, R., Fernandez Vazquez, E., 2020. Entropy-Based solutions for ecological inference problems: A composite estima-
tor. Entropy 22 (7), 781.
Bernardo, J. M., 1984. Monitoring the 1982 spanish socialist victory: A bayesian analysis. Journal of the American Statistical Associ-
ation 79 (387), 510–515.
Bernardo, J. M., 1994. Bayesian Theory. Wiley.
Bernstein, R., 1995. Style Investing. John Wiley & Sons, New York.
Bertsimas, D., Pachamanova, D., 2008. Robust multiperiod portfolio management in the presence of transaction costs. Computers &
Operations Research 35 (1), 3–17.
Bessa, R., Möhrlen, C., Fundel, V., Siefert, M., Browell, J., Haglund El Gaidi, S., Hodge, B.-M., Cali, U., Kariniotakis, G., 2017. Towards
improved understanding of the applicability of uncertainty forecasts in the electric power industry. Energies 10 (9).
Bessa, R. J., Miranda, V., Botterud, A., Zhou, Z., Wang, J., 2012. Time-adaptive quantile-copula for wind power probabilistic forecast-
ing. Renewable Energy 40 (1), 29–39.
Besse, P., Cardot, H., Stephenson, D., 2000. Autoregressive forecasting of some functional climatic variations. Scandinavian Journal
of Statistics 27 (4), 673–687.
Beyaztas, U., Shang, H. L., 2019. Forecasting functional time series using weighted likelihood methodology. Journal of Statistical
Computation and Simulation 89 (16), 3046–3060.
Bianchi, L., Jarrett, J., Hanumara, R. C., 1998. Improving forecasting for telemarketing centers by ARIMA modeling with intervention.
International Journal of Forecasting 14 (4), 497–504.
Bianchi, L., Jarrett, J. E., Hanumara, R. C., 1993. Forecasting incoming calls to telemarketing centers. The Journal of Business Fore-
casting 12 (2), 3.
Bianco, A. M., Garcı́a Ben, M., Martı́nez, E. J., Yohai, V. J., 2001. Outlier detection in regression models with ARIMA errors using
robust estimates. Journal of Forecasting 20 (8), 565–579.
Bickel, J. E., 2007. Some comparisons among quadratic, spherical, and logarithmic scoring rules. Decision Analysis 4 (2), 49–65.
Bickel, P. J., Freedman, D. A., 1981. Some asymptotic theory for the bootstrap. The Annals of Statistics, 1196–1217.
Bielecki, T. R., Rutkowski, M., 2013. Credit risk: modeling, valuation and hedging. Springer Science & Business Media.
Biemer, P. P., 2010. Total survey error: Design, implementation, and evaluation. Public Opinion Quarterly 74 (5), 817–848.
Billio, M., Casarin, R., Ravazzolo, F., van Dijk, H. K., 2013. Time-varying combinations of predictive densities using nonlinear filter-
ing. Journal of Econometrics 177 (2), 213–232.
Binder, C. C., 2017. Measuring uncertainty based on rounding: New method and application to inflation expectations. Journal of
Monetary Economics 90 (C), 1–12.
Bishop, C. M., 2006. Pattern recognition and machine learning. Springer, New York, N.Y.
Bjerknes, V., 1904. Das problem der wettervorhersage, betrachtet vom standpunkte der mechanik und der physik. Meteorologische
Zeitschrift 21, 1–7.
Blanchard, O. J., Kahn, C. M., 1980. The solution of linear difference models under rational expectations. Econometrica 48 (5), 1305–
1311.
Bo, R., Li, F., 2012. Probabilistic LMP forecasting under AC optimal power flow framework: Theory and applications. Electric Power
Systems Research 88, 16–24.

182
Boccara, N., 2004. Modeling complex systems. Springer-Verlag, New York.
Boje, D. M., Murnighan, J. K., 1982. Group confidence pressures in iterative decisions. Management Science 28 (10), 1187–1196.
Bojer, C. S., Meldgaard, J. P., 2020. Kaggle’s forecasting competitions: An overlooked learning opportunity. International Journal of
Forecasting.
Bolger, F., Harvey, N., 1993. Context-sensitive heuristics in statistical reasoning. The Quarterly Journal of Experimental Psychology
Section A 46 (4), 779–811.
Bollerslev, T., 1986. Generalized autoregressive conditional heteroskedasticity. Journal of Econometrics 31 (3), 307–327.
Bollerslev, T., 1987. A Conditionally Heteroskedastic Time Series Model for Speculative Prices and Rates of Return. The Review of
Economics and Statistics 69 (3), 542–547.
Bollerslev, T., 1990. Modelling the Coherence in Short-Run Nominal Exchange Rates: A Multivariate Generalized Arch Model. The
Review of Economics and Statistics 72 (3), 498–505.
Bonaccio, S., Dalal, R. S., 2006. Advice taking and decision-making: An integrative literature review, and implications for the organi-
zational sciences. Organizational Behavior and Human Decision Processes 101 (2), 127–151.
Bonaldo, D., 1991. Competizione tra prodotti farmaceutici: strumenti di previsione. Master’s thesis, University of Padua.
Boneva, L., Fawcett, N., Masolo, R. M., Waldron, M., 2019. Forecasting the UK economy: Alternative forecasting methodologies and
the role of off-model information. International Journal of Forecasting 35 (1), 100–120.
Bonham, C., Cohen, R., 2001. To aggregate, pool, or neither: Testing the rational expectations hypothesis using survey data. Journal
of Business & Economic Statistics 190, 278–291.
Booij, N., Ris, R. C., Holthuijsen, L. H., 1999. A third-generation wave model for coastal regions 1. Model description and validation.
Journal of Geophysical Research: Oceans 104 (C4), 7649–7666.
Boone, T., Ganeshan, R., 2008. The value of information sharing in the retail supply chain: Two case studies. Foresight: The Interna-
tional Journal of Applied Forecasting 9, 12–17.
Booth, H., Tickle, L., 2008. Mortality modelling and forecasting: A review of methods. Annals of Actuarial Science 3 (1-2), 3–43.
Bordalo, P., Gennaioli, N., Ma, Y., Shleifer, A., 2018. Over-reaction in Macroeconomic Expectations. NBER Working Papers 24932,
National Bureau of Economic Research, Inc.
Bordignon, S., Bunn, D. W., Lisi, F., Nan, F., 2013. Combining day-ahead forecasts for british electricity prices. Energy Economics 35,
88–103.
Bordley, R. F., 1982. The combination of forecasts: a Bayesian approach. Journal of the Operational Research Society 33 (2), 171–174.
Bork, L., Møller, S. V., 2015. Forecasting house prices in the 50 states using dynamic model averaging and dynamic model selection.
International Journal of Forecasting 31 (1), 63–78.
Bosq, D., 2000. Linear Processes in Function Spaces. Lecture Notes in Statistics, New York.
Bosq, D., Blanke, D., 2007. Inference and Prediction in Large Dimensions. John Wiley & Sons, West Sussex, England.
Boswijk, H. P., Franses, P. H., 1996. Unit roots in periodic autoregressions. Journal of Time Series Analysis 17 (3), 221–245.
Boswijk, H. P., Franses, P. H., Haldrup, N., 1997. Multiple unit roots in periodic autoregression. Journal of Econometrics 80 (1),
167–193.
Bourdeau, M., qiang Zhai, X., Nefzaoui, E., Guo, X., Chatellier, P., 2019. Modeling and forecasting building energy consumption: A
review of data-driven techniques. Sustainable Cities and Society 48, 101533.
Box, George, E. P., Jenkins, Gwilym, 1976. Time Series Analysis Forecasting and Control. Holden-Day, San Francisco, CA.
Box, G. E. P., Jenkins, G. M., Reinsel, G. C., 2008. Time Series Analysis: Forecasting and Control, 4th Edition. Wiley, New Jersey.
Boylan, J., Syntetos, A., 2006. Accuracy and accuracy implication metrics for intermittent demand. Foresight: The International
Journal of Applied Forecasting 4, 39–42.
Boylan, J. E., Babai, M. Z., 2016. On the performance of overlapping and non-overlapping temporal demand aggregation approaches.
International Journal of Production Economics 181, 136–144.
Boylan, J. E., Syntetos, A. A., 2003. Intermittent demand forecasting: size-interval methods based on averaging and smoothing. In:
Frangos, C. C. (Ed.), Proceedings of the International Conference on Quantitative Methods in Industry and Commerce. Techno-
logical Educational Institute, Athens, pp. 87–96.
Boylan, J. E., Syntetos, A. A., 2021. Intermittent Demand Forecasting - Context, Methods and Applications. Wiley.
Boylan, J. E., Syntetos, A. A., Karakostas, G. C., 2008. Classification for forecasting and stock control: a case study. Journal of the
Operational Research Society 59 (4), 473–481.
Bozkurt, Ö. Ö., Biricik, G., Tayşi, Z. C., 2017. Artificial neural network and SARIMA based models for power load forecasting in
turkish electricity market. PloS One 12 (4), e0175915.
Brandt, M. W., Jones, C. S., 2006. Volatility forecasting with range-based EGARCH models. Journal of Business & Economic Statistics

183
 24 (4), 470–486.
Braumoeller, B. F., 2019. Only the Dead: The Persistence of War in the Modern Age. Oxford University Press.
Brehmer, J., Gneiting, T., 2020. Scoring interval forecasts: Equal-tailed, shortest, and modal interval. arXiv:2007.05709.
Breiman, L., 1996. Bagging predictors. Machine Learning 24 (2), 123–140.
Breiman, L., 2001. Random forests. Machine Learning 45 (1), 5–32.
Brennan, J., 2020. Can novices trust themselves to choose trustworthy experts? reasons for (reserved) optimism. Social Epistemology
34 (3), 227–240.
Brier, G. W., 1950. Verification of forecasts expressed in terms of probability. Monthly Weather Review 78 (1), 1–3.
Brighton, H., Gigerenzer, G., 2015. The bias bias. Journal of Business Research 68 (8), 1772–1784.
Broer, T., Kohlhas, A., 2018. Forecaster (Mis-)Behavior. CEPR Discussion Papers 12898, C.E.P.R. Discussion Papers.
Brown, A., Reade, J. J., Vaughan Williams, L., 2019. When are prediction market prices most informative? International Journal of
Forecasting 35 (1), 420–428.
Brown, G., Wyatt, J., Harris, R., Yao, X., 2005a. Diversity creation methods: a survey and categorisation. Information Fusion 6 (1),
5–20.
Brown, L., Gans, N., Mandelbaum, A., Sakov, A., Shen, H., Zeltyn, S., Zhao, L., 2005b. Statistical analysis of a telephone call center:
A queueing-science perspective. Journal of the American Statistical Association 100 (469), 36–50.
Brown, L., Moshiri, S., 2004. Unemployment variation over the business cycles: A comparison of forecasting models. Journal of
Forecasting 23 (7), 497–511.
Brunetti, C., Lildholdt, P. M., 2002. Return-based and range-based (co)variance estimation - with an application to foreign exchange
markets. SSRN:296875.
Buchanan, B. G., 2019. Artificial intelligence in finance. The Alan Turing Institute, London.
Buckle, H. T., 1858. History of Civilization in England. Vol. 1. John W. Parker and Son.
Budescu, D. V., Wallsten, T. S., 1985. Consistency in interpretation of probabilistic phrases. Organizational Behavior and Human
Decision Processes 36 (3), 391–405.
Bühlmann, P., 1997. Sieve bootstrap for time series. Bernoulli 3 (2), 123–148.
Buizza, R., 2018. Ensemble forecasting and the need for calibration. In: Statistical postprocessing of ensemble forecasts. Elsevier, pp.
15–48.
Bunea, A., Della Posta, P., Guidolin, M., Manfredi, P., 2020. What do adoption patterns of solar panels observed so far tell about
governments’ incentive? Insights from diffusion models. Technological Forecasting and Social Change 160, 120240.
Bunn, D. W., 1975. A Bayesian approach to the linear combination of forecasts. Journal of the Operational Research Society 26 (2),
325–329.
Bunn, D. W., Salo, A. A., 1993. Forecasting with scenarios. European Journal of Operational Research 68 (3), 291–303.
Bureau of Transportation Statistics, 2020. Reporting carrier on-time performance (1987 - present). Accessed on 2020-09-09.
Burgman, M. A., 2016. Trusting Judgements: How to Get the Best out of Experts. Cambridge University Press.
Burman, P., Chow, E., Nolan, D., 1994. A cross-validatory method for dependent data. Biometrika 81 (2), 351–358.
Burridge, P., Robert Taylor, A., 2006. Additive outlier detection via extreme-value theory. Journal of Time Series Analysis 27 (5),
685–701.
Burton, J. W., Stein, M., Jensen, T. B., 2020. A systematic review of algorithm aversion in augmented decision making. Journal of
Behavioral Decision Making 33 (2), 220–239.
Busetti, F., Marcucci, J., 2013. Comparing forecast accuracy: a Monte Carlo investigation. International Journal of Forecasting 29 (1),
13–27.
Butler, D., Butler, R., Eakins, J., 2020. Expert performance and crowd wisdom: Evidence from english premier league predictions.
European Journal of Operational Research 288, 170–182.
Buys-Ballot, C. H. D., 1847. Les changements périodiques de temperature. Utrecht: Kemink et Fils.
Byrne, J. P., Fazio, G., Fiess, N., 2013. Primary commodity prices: Co-movements, common factors and fundamentals. Journal of
Development Economics 101, 16–26.
Byrne, J. P., Korobilis, D., Ribeiro, P. J., 2016. Exchange rate predictability in a changing world. Journal of International Money and
Finance 62, 1–24.
Ca’ Zorzi, M., Cap, A., Mijakovic, A., Rubaszek, M., 2020. The predictive power of equilibrium exchange rate models. Working Paper
Series 2358, European Central Bank.
Ca’ Zorzi, M., Kolasa, M., Rubaszek, M., 2017. Exchange rate forecasting with DSGE models. Journal of International Economics
107 (C), 127–146.

184
Ca’ Zorzi, M., Muck, J., Rubaszek, M., 2016. Real exchange rate forecasting and PPP: This time the random walk loses. Open
Economies Review 27 (3), 585–609.
Ca’ Zorzi, M., Rubaszek, M., 2020. Exchange rate forecasting on a napkin. Journal of International Money and Finance 104, 102168.
Cai, J., 1994. A Markov model of Switching-Regime ARCH. Journal of Business & Economic Statistics 12 (3), 309–316.
Cairns, A. J. G., Blake, D., Dowd, K., Coughlan, G. D., Epstein, D., Ong, A., Balevich, I., 2009. A quantitative comparison of stochastic
mortality models using data from England and Wales and the United States. North American Actuarial Journal 13 (1), 1–35.
Calvo, E., Escolar, M., 2003. The local voter: A geographically weighted approach to ecological inference. American Journal of Political
Science 47 (1), 189–204.
Campbell, J. Y., Thompson, S. B., 2008. Predicting excess stock returns out of sample: Can anything beat the historical average?
Review of Financial Studies 21 (4), 1509–1531.
Cannon, E., Tonks, I., 2008. Annuity Markets. Oxford University Press, Oxford.
Cardani, R., Paccagnini, A., Villa, S., 2015. Forecasting in a DSGE Model with Banking Intermediation: Evidence from the US.
Working Paper 292, University of Milano-Bicocca, Department of Economics.
Cardani, R., Paccagnini, A., Villa, S., 2019. Forecasting with instabilities: An application to DSGE models with financial frictions.
Journal of Macroeconomics 61 (C), 103133.
Carlstein, E., 1990. Resampling techniques for stationary time-series: some recent developments. Tech. rep., North Carolina State
University, Department of Statistics.
Carmo, J. L., Rodrigues, A. J., 2004. Adaptive forecasting of irregular demand processes. Engineering Applications of Artificial
Intelligence 17 (2), 137–143.
Carnevale, C., Angelis, E. D., Finzi, G., Turrini, E., Volta, M., 2020. Application of data fusion techniques to improve air quality
forecast: A case study in the northern Italy. Atmosphere 11 (3).
Carnevale, C., Finzi, G., Pederzoli, A., Turrini, E., Volta, M., 2018. An integrated data-driven/data assimilation approach for the
forecast of PM10 levels in northern Italy. In: Mensink, C., Kallos, G. (Eds.), Air Pollution Modeling and its Application XXV.
Springer International Publishing, pp. 225–229.
Carnevale, C., Finzi, G., Pisoni, E., Volta, M., 2016. Lazy learning based surrogate models for air quality planning. Environmental
Modelling and Software 83, 47–57.
Carriero, A., Clements, M. P., Galvão, A. B., 2015. Forecasting with Bayesian multivariate vintage-based VARs. International Journal
of Forecasting 31 (3), 757–768.
Carroll, R., 2003. The Skeptic’s Dictionary: A Collection of Strange Beliefs, Amusing Deceptions, and Dangerous Delusions. Wiley.
Carvalho, C. M., Polson, N. G., Scott, J. G., 2010. The horseshoe estimator for sparse signals. Biometrika 97 (2), 465–480.
Carvalho, T. P., Soares, F. A. A. M. N., Vita, R., Francisco, R. d. P., Basto, J. P., Alcalá, S. G. S., 2019. A systematic literature review of
machine learning methods applied to predictive maintenance. Computers & Industrial Engineering 137, 106024.
Casals, J., Garcia-Hiernaux, A., Jerez, M., Sotoca, S., Trindade, A., 2016. State-Space Methods for Time Series Analysis: Theory,
Applications and Software. Chapman-Hall / CRC Press.
Castle, J., Doornik, J., Hendry, D., Pretis, F., 2015a. Detecting Location Shifts during Model Selection by Step-Indicator Saturation.
Econometrics 3 (2), 240–264.
Castle, J. L., Clements, M. P., Hendry, D. F., 2015b. Robust Approaches to Forecasting. International Journal of Forecasting 31 (1),
99–112.
Castle, J. L., Doornik, J. A., Hendry, D. F., 2018. Selecting a model for forecasting. Working paper, Economics Department, Oxford
University.
Castle, J. L., Doornik, J. A., Hendry, D. F., 2020a. Modelling non-stationary ‘big data’. International Journal of Forecasting.
Castle, J. L., Doornik, J. A., Hendry, D. F., 2020b. Robust discovery of regression models. Working paper 2020-W04, Nuffield College,
Oxford University.
Castle, J. L., Doornik, J. A., Hendry, D. F., Pretis, F., 2015c. Detecting location shifts during model selection by step-indicator satura-
tion. Econometrics 3(2), 240–264.
Castle, J. L., Fawcett, N. W., Hendry, D. F., 2010. Forecasting with equilibrium-correction models during structural breaks. Journal of
Econometrics 158 (1), 25–36.
Castle, J. L., Hendry, D. F., 2020a. Climate Econometrics: An Overview. Foundations and Trends in Econometrics 10, 145–322.
Castle, J. L., Hendry, D. F., 2020b. Identifying the causal role of CO2 during the Ice Ages. Discussion paper 898, Economics Depart-
ment, Oxford University.
Castle, J. L., Hendry, D. F., Kitov, O. K., 2017. Forecasting and nowcasting macroeconomic variables: A methodological overview.
Handbook on rapid estimates, UN and Eurostat, chapter 3.

185
Castle, J. L., Hendry, D. F., Martinez, A. B., 2020c. The paradox of stagnant real wages yet rising ‘living standards’ in the UK. Tech.
rep., VoxEU.
Catalán, B., Trı́vez, F. J., 2007. Forecasting volatility in GARCH models with additive outliers. Quantitative Finance 7 (6), 591–596.
Cavalcante, L., Bessa, R. J., Reis, M., Browell, J., 2016. Lasso vector autoregression structures for very short-term wind power fore-
casting. Wind Energy 20, 657–675.
Cazelles, B., Chavez, M., McMichael, A. J., Hales, S., 2005. Nonstationary influence of el nino on the synchronous dengue epidemics
in Thailand. PLoS Medicine 2 (4), e106.
Cederman, L.-E., 2003. Modeling the size of wars: From billiard balls to sandpiles. The American Political Science Review 97 (1),
135–150.
Ceron, A., Curini, L., Iacus, S. M., 2016. Politics and Big Data: Nowcasting and Forecasting Elections with Social Media. Routledge.
Chae, Y. T., Horesh, R., Hwang, Y., Lee, Y. M., 2016. Artificial neural network model for forecasting sub-hourly electricity usage in
commercial buildings. Energy and Buildings 111, 184–194.
Chakraborty, T., Chattopadhyay, S., Ghosh, I., 2019. Forecasting dengue epidemics using a hybrid methodology. Physica A: Statistical
Mechanics and its Applications 527, 121266.
Chan, C. K., Kingsman, B. G., Wong, H., 1999. The value of combining forecasts in inventory management–a case study in banking.
European Journal of Operational Research 117 (2), 199–210.
Chan, F., Pauwels, L. L., 2018. Some theoretical results on forecast combinations. International Journal of Forecasting 34 (1), 64–74.
Chan, J. S., Lam, C. P., Yu, P. L., Choy, S. T., Chen, C. W., 2012. A Bayesian conditional autoregressive geometric process model for
range data. Computational Statistics and Data Analysis 56 (11), 3006–3019.
Chan, K. S., Tong, H., 1986. On estimating thresholds in autoregressive models. Journal of Time Series Analysis 7 (3), 179–190.
Chang, Y., Kim, C. S., Park, J., 2016. Nonstationarity in time series of state densities. Journal of Econometrics 192 (1), 152–167.
Chaouch, M., 2014. Clustering-Based improvement of nonparametric functional time series forecasting: Application to Intra-Day
Household-Level load curves. IEEE Transactions on Smart Grid 5 (1), 411–419.
Chase, C., 2021. Assisted demand planning using machine learning. In: Gilliland, M., Tashman, L., Sglavo, U. (Eds.), Business
Forecasting: The Emerging Role of Artificial Intelligence and Machine Learning. Wiley.
Chatfield, C., 1986. Simple is best? International Journal of Forecasting 2 (4), 401–402.
Chatziantoniou, I., Degiannakis, S., Eeckels, B., Filis, G., 2016. Forecasting tourist arrivals using origin country macroeconomics.
Applied Economics 48 (27), 2571–2585.
Chavez-Demoulin, V., Davison, A. C., McNeil, A. J., 2005. Estimating value-at-risk: a point process approach. Quantitative Finance
5 (2), 227–234.
Checchi, F., Roberts, L., 2005. Interpreting and using mortality data in humanitarian emergencies. Humanitarian Practice Network
52.
Chen, C., Liu, L.-M., 1993a. Forecasting time series with outliers. Journal of Forecasting 12 (1), 13–35.
Chen, C., Liu, L.-M., 1993b. Joint estimation of model parameters and outlier effects in time series. Journal of the American Statistical
Association 88 (421), 284–297.
Chen, C. W., Gerlach, R., Hwang, B. B., McAleer, M., 2012. Forecasting Value-at-Risk using nonlinear regression quantiles and the
intra-day range. International Journal of Forecasting 28 (3), 557–574.
Chen, C. W., Gerlach, R., Lin, E. M., 2008. Volatility forecasting using threshold heteroskedastic models of the intra-day range.
Computational Statistics and Data Analysis 52 (6), 2990–3010.
Chen, C. W. S., Chiang, T. C., So, M. K. P., 2003. Asymmetrical reaction to US stock-return news: evidence from major stock markets
based on a double-threshold model. Journal of Economics and Business 55 (5), 487–502.
Chen, C. W. S., So, M. K. P., 2006. On a threshold heteroscedastic model. International Journal of Forecasting 22 (1), 73–89.
Chen, J., Li, K., Rong, H., Bilal, K., Li, K., Philip, S. Y., 2019a. A periodicity-based parallel time series prediction algorithm in cloud
computing environments. Information Sciences 496, 506–537.
Chen, M.-F., Wang, R.-H., Hung, S.-L., 2015. Predicting health-promoting self-care behaviors in people with pre-diabetes by applying
Bandura social learning theory. Applied Nursing Research 28 (4), 299–304.
Chen, R., 1995. Threshold variable selection in open-loop threshold autoregressive models. Journal of Time Series Analysis 16 (5),
461–481.
Chen, T., Guestrin, C., 2016. Xgboost: A scalable tree boosting system. In: Proceedings of the 22nd ACM SIGKDD International
Conference on Knowledge Discovery and Data Mining. ACM, pp. 785–794.
Chen, Y., Chu, C.-H., Mullen, T., Pennock, D. M., 2005. Information markets vs. opinion pools: an empirical comparison. In: Pro-
ceedings of the 6th ACM conference on Electronic commerce. EC ’05. Association for Computing Machinery, New York, NY, USA,

186
 pp. 58–67.
Chen, Y., Kang, Y., Chen, Y., Wang, Z., 2020. Probabilistic forecasting with temporal convolutional neural network. Neurocomputing
399, 491–501.
Chen, Y., Marron, J. S., Zhang, J., 2019b. Modeling seasonality and serial dependence of electricity price curves with warping func-
tional autoregressive dynamics. The Annals of Applied Statistics 13 (3), 1590–1616.
Cheng, C., Yu, L., Chen, L. J., 2012. Structural nonlinear damage detection based on ARMA-GARCH model. Applied Mechanics and
Materials 204-208, 2891–2896.
Cheng, G., Yang, Y., 2015. Forecast combination with outlier protection. International Journal of Forecasting 31 (2), 223–237.
Cheung, Y.-W., Chinn, M. D., Pascual, A. G., 2005. Empirical exchange rate models of the nineties: Are any fit to survive? Journal of
International Money and Finance 24 (7), 1150–1175.
Cheung, Y.-W., Chinn, M. D., Pascual, A. G., Zhang, Y., 2019. Exchange rate prediction redux: New models, new data, new currencies.
Journal of International Money and Finance 95, 332–362.
Chew, V., 1968. Simultaneous prediction intervals. Technometrics 10 (2), 323–330.
Chiang, M. H., Wang, L. M., 2011. Volatility contagion: A range-based volatility approach. Journal of Econometrics 165 (2), 175–189.
Chicco, G., Cocina, V., Di Leo, P., Spertino, F., Massi Pavan, A., 2015. Error assessment of solar irradiance forecasts and AC power
from energy conversion model in Grid-Connected photovoltaic systems. Energies 9 (1), 8.
Chinco, A., Clark-Joseph, A. D., Ye, M., 2019. Sparse signals in the cross-section of returns. Journal of Finance 74 (1), 449–492.
Chiroma, H., Abdulkareem, S., Herawan, T., 2015. Evolutionary Neural Network model for West Texas Intermediate crude oil price
prediction. Applied Energy 142, 266–273.
Choi, E., Özer, Ö., Zheng, Y., 2020. Network trust and trust behaviors among executives in supply chain interactions. Management
Science.
Chong, Y. Y., Hendry, D. F., 1986. Econometric evaluation of linear macro-economic models. The Review of Economic Studies 53 (4),
671–690.
Chou, R. Y., Cai, Y., 2009. Range-based multivariate volatility model with double smooth transition in conditional correlation. Global
Finance Journal 20 (2), 137–152.
Chou, R. Y., Chou, H., Liu, N., 2015. Range volatility: A review of models and empirical studies. In: Lee, C. F., Lee, J. C. (Eds.),
Handbook of Financial Econometrics and Statistics. Springer New York, pp. 2029–2050.
Chou, R. Y., Liu, N., 2010. The economic value of volatility timing using a range-based volatility model. Journal of Economic Dynam-
ics and Control 34 (11), 2288–2301.
Chou, R. Y., Wu, C. C., Liu, N., 2009. Forecasting time-varying covariance with a range-based dynamic conditional correlation model.
Review of Quantitative Finance and Accounting 33 (4), 327–345.
Chou, R. Y.-T., 2005. Forecasting Financial Volatilities with Extreme Values: The Conditional Autoregressive Range (CARR) Model.
Journal of Money, Credit, and Banking 37 (3), 561–582.
Choudhury, A., Urena, E., 2020. Forecasting hourly emergency department arrival using time series analysis. British Journal of
Healthcare Management 26 (1), 34–43.
Christ, M., Braun, N., Neuffer, J., Kempa-Liehr, A. W., 2018. Time Series FeatuRe Extraction on basis of Scalable Hypothesis tests
(tsfresh – a Python package). Neurocomputing 307, 72–77.
Christensen, T., Hurn, S., Lindsay, K., 2009. It never rains but it pours: Modeling the persistence of spikes in electricity prices. Energy
Journal 30 (1), 25–48.
Christensen, T. M., Hurn, A. S., Lindsay, K. A., 2012. Forecasting spikes in electricity prices. International Journal of Forecasting
28 (2), 400–411.
Christiano, L. J., Eichenbaum, M. S., Trabandt, M., 2018. On DSGE models. Journal of Economic Perspectives 32 (3), 113–40.
Christoffersen, P., Langlois, H., 2013. The joint dynamics of equity market factors. Journal of Financial and Quantitative Analysis
48 (5), 1371–1404.
Chung, C., Niu, S.-C., Sriskandarajah, C., 2012. A sales forecast model for short-life-cycle products: New releases at blockbuster.
Production and Operations Management 21 (5), 851–873.
Chung, H., Kiley, M. T., Laforte, J.-P., 2010. Documentation of the Estimated, Dynamic, Optimization-based (EDO) model of the U.S.
economy: 2010 version. Finance and Economics Discussion Series 2010-29, Board of Governors of the Federal Reserve System
(U.S.).
Cirillo, P., Taleb, N. N., 2016a. Expected shortfall estimation for apparently infinite-mean models of operational risk. Quantitative
Finance 16 (10), 1485–1494.
Cirillo, P., Taleb, N. N., 2016b. On the statistical properties and tail risk of violent conflicts. Physica A: Statistical Mechanics and its

187
 Applications 452, 29–45.
Cirillo, P., Taleb, N. N., 2019. The decline of violent conflicts: What do the data really say? In: Toje, A., Steen, B. N. V. (Eds.), The
Causes of Peace: What We Know. The Causes of Peace: What We Know, pp. 57–86.
Claeskens, G., Magnus, J. R., Vasnev, A. L., Wang, W., 2016. The forecast combination puzzle: A simple theoretical explanation.
International Journal of Forecasting 32 (3), 754–762.
Clark, D. A., 1990. Verbal uncertainty expressions: A critical review of two decades of research. Current Psychology 9 (3), 203–235.
Clark, T., McCracken, M., 2013. Advances in forecast evaluation. In: Handbook of Economic Forecasting. Vol. 2. Elsevier, pp. 1107–
1201.
Clark, T., West, K., 2006. Using out-of-sample mean squared prediction errors to test the martingale difference hypothesis. Journal of
Econometrics 135 (1-2), 155–186.
Clark, T. E., McCracken, M. W., 2001. Tests of equal forecast accuracy and encompassing for nested models. Journal of Econometrics
105 (1), 85–110.
Clark, T. E., McCracken, M. W., 2009. Tests of equal predictive ability with real-time data. Journal of Business & Economic Statistics
27 (4), 441–454.
Clauset, A., 2018. Trends and fluctuations in the severity of interstate wars. Science Advances 4 (2), eaao3580.
Clauset, A., Gleditsch, K. S., 2018. Trends in conflicts: What do we know and what can we know? In: Gheciu, A., Wohlforth, W. C.
(Eds.), The Oxford Handbook of International Security. Oxford University Press.
Cleave, N., Brown, P. J., Payne, C. D., 1995. Evaluation of methods for ecological inference. Journal of the Royal Statistical Society,
Series A, 158 (1), 55–72.
Clemen, R. T., 1989. Combining forecasts: A review and annotated bibliography. International Journal of Forecasting 5 (4), 559–583.
Clemen, R. T., 2008. Comment on cooke’s classical method. Reliability Engineering & System Safety 93 (5), 760–765.
Clemen, R. T., Winkler, R. L., 1986. Combining economic forecasts. Journal of Business & Economic Statistics 4 (1), 39–46.
Clements, A. E., Herrera, R., Hurn, A. S., 2015. Modelling interregional links in electricity price spikes. Energy Economics 51, 383–
393.
Clements, M. P., 2009. Internal consistency of survey respondents’ forecasts: Evidence based on the Survey of Professional Forecasters.
In: Castle, J. L., Shephard, N. (Eds.), The Methodology and Practice of Econometrics. A Festschrift in Honour of David F. Hendry.
Chapter 8. Oxford University Press, Oxford, pp. 206–226.
Clements, M. P., 2010. Explanations of the Inconsistencies in Survey Respondents Forecasts. European Economic Review 54 (4),
536–549.
Clements, M. P., 2011. An empirical investigation of the effects of rounding on the SPF probabilities of decline and output growth
histograms. Journal of Money, Credit and Banking 43 (1), 207–220.
Clements, M. P., 2014a. Forecast Uncertainty - Ex Ante and Ex Post: US Inflation and Output Growth. Journal of Business & Economic
Statistics 32 (2), 206–216.
Clements, M. P., 2014b. US inflation expectations and heterogeneous loss functions, 1968–2010. Journal of Forecasting 33 (1), 1–14.
Clements, M. P., 2017. Assessing macro uncertainty in real-time when data are subject to revision. Journal of Business & Economic
Statistics 35 (3), 420–433.
Clements, M. P., 2018. Are macroeconomic density forecasts informative? International Journal of Forecasting 34, 181–198.
Clements, M. P., 2019. Macroeconomic Survey Expectations. Palgrave Texts in Econometrics. Palgrave Macmillan.
Clements, M. P., Galvão, A. B., 2012. Improving real-time estimates of output gaps and inflation trends with multiple-vintage VAR
models. Journal of Business & Economic Statistics 30 (4), 554–562.
Clements, M. P., Galvão, A. B., 2013a. Forecasting with vector autoregressive models of data vintages: US output growth and inflation.
International Journal of Forecasting 29 (4), 698–714.
Clements, M. P., Galvão, A. B., 2013b. Real-time forecasting of inflation and output growth with autoregressive models in the presence
of data revisions. Journal of Applied Econometrics 28 (3), 458–477.
Clements, M. P., Galvão, A. B., 2017. Data revisions and real-time probabilistic forecasting of macroeconomic variables. Discussion
Paper ICM-2017-01, ICMA, Henley Business School, Reading.
Clements, M. P., Galvão, A. B., 2019. Data revisions and real-time forecasting. The Oxford Research Encyclopedia of Economics and
Finance.
Clements, M. P., Galvão, A. B. C., 2003. Testing the expectations theory of the term structure of interest rates in threshold models.
Macroeconomic Dynamics 7 (4), 567–585.
Clements, M. P., Harvey, D. I., 2011. Combining probability forecasts. International Journal of Forecasting 27 (2), 208–223.
Clements, M. P., Hendry, D. F., 1998. Forecasting Economic Time Series. Cambridge University Press.

188
Clements, M. P., Hendry, D. F., 1999. Forecasting Non-stationary Economic Time Series. Zeuthen Lecture Book Series. Cambridge,
MA: MIT Press.
Clements, M. P., Hendry, D. F., 2005. Evaluating a model by forecast performance. Oxford Bulletin of Economics and Statistics 67,
931–956.
Cleveland, R. B., Cleveland, W. S., McRae, J. E., Terpenning, I., 1990. STL: A seasonal-trend decomposition procedure based on Loess.
Journal of Official Statistics 6 (1), 3–73.
Clottey, T., Benton, Jr., W. C., Srivastava, R., 2012. Forecasting product returns for remanufacturing operations. Decision Sciences
43 (4), 589–614.
Coates, D., Humphreys, B. R., 1999. The growth effects of sport franchises, stadia, and arenas. Journal of Policy Analysis and Man-
agement 18 (4), 601–624.
Coccia, G., 2011. Analysis and developments of uncertainty processors for real time flood forecasting. Ph.D. thesis, Alma Mater
Studiorum University of Bologna.
Coccia, G., Todini, E., 2011. Recent developments in predictive uncertainty assessment based on the model conditional processor
approach. Hydrology and Earth System Sciences 15 (10), 3253–3274.
Coibion, O., Gorodnichenko, Y., 2012. What can survey forecasts tell us about information rigidities? Journal of Political Economy
120 (1), 116–159.
Coibion, O., Gorodnichenko, Y., 2015. Information Rigidity and the Expectations Formation Process: A Simple Framework and New
Facts. American Economic Review 105 (8), 2644–78.
Collopy, F., Armstrong, J. S., 1992. Rule-based forecasting: development and validation of an expert systems approach to combining
time series extrapolations. Management Science 38 (10), 1394–1414.
Commandeur, J. J. F., Koopman, S. J., Ooms, M., 2011. Statistical software for state space methods. Journal of Statistical Software
41 (1), 1–18.
Consolo, A., Favero, C., Paccagnini, A., 2009. On the statistical identification of DSGE models. Journal of Econometrics 150 (1),
99–115.
Continuous Mortality Investigation, 2020. The CMI Mortality Projections Model, CMI 2019. Working paper, Institute of Actuaries
and Faculty of Actuaries, London.
Cook, S., Thomas, C., 2003. An alternative approach to examining the ripple effect in U.K. house prices. Applied Economics Letters
10 (13), 849–851.
Cooke, R. M., 1991. Experts in Uncertainty: Opinion and Subjective Probability in Science. Oxford University Press.
Copeland, M. T., 1915. Statistical indices of business conditions. The Quarterly Journal of Economics 29 (3), 522–562.
Corani, G., 2005. Air quality prediction in milan: Feed-forward neural networks, pruned neural networks and lazy learning. Ecolog-
ical Modeling 185, 513–529.
Cordeiro, C., Neves, M., 2006. The bootstrap methodology in time series forecasting. In: Black, J., White, A. (Eds.), Proceedings of
CompStat2006. Springer Verlag, pp. 1067–1073.
Cordeiro, C., Neves, M., 2009. Forecasting time series with BOOT.EXPOS procedure. REVSTAT-Statistical Journal 7 (2), 135–149.
Cordeiro, C., Neves, M. M., 2010. Boot.EXPOS in nngc competition. In: The 2010 International Joint Conference on Neural Networks
(IJCNN). IEEE, pp. 1–7.
Cordeiro, C., Neves, M. M., 2013. Predicting and treating missing data with Boot.EXPOS. In: Advances in Regression, Survival
Analysis, Extreme Values, Markov Processes and Other Statistical Applications. Springer, pp. 131–138.
Cordeiro, C., Neves, M. M., 2014. Forecast intervals with Boot.EXPOS. In: New Advances in Statistical Modeling and Applications.
Springer, pp. 249–256.
Corominas, A., Lusa, A., Dolors Calvet, M., 2015. Computing voter transitions: The elections for the Catalan parliament, from 2010
to 2012. Journal of Industrial Engineering and Management 8 (1), 122–136.
Corradi, V., Swanson, N. R., Olivetti, C., 2001. Predictive ability with cointegrated variables. Journal of Econometrics 104 (2), 315–
358.
Couharde, C., Delatte, A.-L., Grekou, C., Mignon, V., Morvillier, F., 2018. EQCHANGE: A world database on actual and equilibrium
effective exchange rates. International Economics 156, 206–230.
Crawford, T., de Haan, R., Runchey, C., 2008. Longevity risk quantification and management: A review of relevant literature. Working
paper, Society of Actuaries.
Creal, D., Koopman, S. J., Lucas, A., 2013. Generalized autoregressive score models with applications. Journal of Applied Economet-
rics 28, 777–795.
Creal, D. D., Tsay, R. S., 2015. High dimensional dynamic stochastic copula models. Journal of Econometrics 189 (2), 335–345.

189
Croll, J., 1875. Climate and Time in Their Geological Relations, A Theory of Secular Changes of the Earth’s Climate. D. Appleton,
New York.
Crone, S. F., Hibon, M., Nikolopoulos, K., 2011. Advances in forecasting with neural networks? Empirical evidence from the NN3
competition on time series prediction. International Journal of Forecasting 27 (3), 635–660.
Cross, J. L., 2020. Macroeconomic forecasting with large bayesian vars: Global-local priors and the illusion of sparsity. International
Journal of Forecasting 36 (3), 899–916.
Cross, R., Sproull, L., 2004. More than an answer: Information relationships for actionable knowledge. Organization Science 15 (4),
446–462.
Croston, J. D., 1972. Forecasting and stock control for intermittent demands. Operational Research Quarterly 23 (3), 289–303.
Croushore, D., 2006. Forecasting with real-time macroeconomic data. In: Elliott, G., Granger, C., Timmermann, A. (Eds.), Handbook
of Economic Forecasting, Volume 1. Handbook of Economics 24. Elsevier, Horth-Holland, pp. 961–982.
Croushore, D., 2011a. Forecasting with real-time data vintages, chapter 9. In: Clements, M. P., Hendry, D. F. (Eds.), The Oxford
Handbook of Economic Forecasting. Oxford University Press, pp. 247–267.
Croushore, D., 2011b. Frontiers of real-time data analysis. Journal of Economic Literature 49, 72–100.
Croushore, D., Stark, T., 2001. A real-time data set for macroeconomists. Journal of Econometrics 105 (1), 111–130.
Croxson, K., Reade, J. J., 2014. Information and efficiency: goal arrival in soccer. The Economic Journal 124 (575), 62–91.
Cui, R., Zhang, D. J., Bassamboo, A., 2019. Learning from inventory availability information: Evidence from field experiments on
amazon. Management Science 65 (3), 1216–1235.
Cunado, J., De Gracia, F. P., 2005. Oil prices, economic activity and inflation: Evidence for some Asian countries. The Quarterly
Review of Economics and Finance 45 (1), 65–83.
Cunningham, A., Eklund, J., Jeffery, C., Kapetanios, G., Labhard, V., 2009. A state space approach to extracting the signal from
uncertain data. Journal of Business & Economic Statistics 30, 173–180.
Cunningham, C. R., 2006. House price uncertainty, timing of development, and vacant land prices: Evidence for real options in
seattle. Journal of Urban Economics 59 (1), 1–31.
Curran, M., Velic, A., 2019. Real exchange rate persistence and country characteristics: A global analysis. Journal of International
Money and Finance 97, 35–56.
Cybenko, G., 1989. Approximation by superpositions of a sigmoidal function. Mathematics of Control, Signals, and Systems 2 (4),
303–314.
Dagum, E. B., 1988. The X11ARIMA/88 Seasonal Adjustment Method: Foundations and User’s Manual. Statistics Canada, Time
Series Research and Analysis Division.
Dai, Q., Singleton, K., 2003. Term structure dynamics in theory and reality. The Review of Financial Studies 16 (3), 631–678.
Dai, Q., Singleton, K. J., Yang, W., 2007. Regime shifts in a dynamic term structure model of U.S. treasury bond yields. The Review of
Financial Studies 20 (5), 1669–1706.
Dalkey, N. C., 1969. The Delphi method: An experimental study of group opinion. Research Memoranda RM-5888-PR.
Dalla Valle, A., Furlan, C., 2014. Diffusion of nuclear energy in some developing and graduated developing countries. Technological
Forecasting and Social Change 18, 143–153.
Daniel, K., Moskowitz, T. J., 2016. Momentum crashes. Journal of Financial Economics 122 (2), 221–247.
Dantas, T. M., Cyrino Oliveira, F. L., 2018. Improving time series forecasting: An approach combining bootstrap aggregation, clusters
and exponential smoothing. International Journal of Forecasting 34 (4), 748–761.
Dantas, T. M., Cyrino Oliveira, F. L., Varela Repolho, H. M., 2017. Air transportation demand forecast through bagging holt winters
methods. Journal of Air Transport Management 59, 116–123.
Danti, P., Magnani, S., 2017. Effects of the load forecasts mismatch on the optimized schedule of a real small-size smart prosumer.
Energy Procedia 126, 406–413.
Dantzig, G. B., Infanger, G., 1993. Multi-stage stochastic linear programs for portfolio optimization. Annals of Operations Research
45, 59–76.
Das, S., Chen, M., 2007. Yahoo! for Amazon: Sentiment extraction from small talk on the web. Management Science 53 (9), 1375–
1388.
Daskalaki, C., Kostakis, A., Skiadopoulos, G., 2014. Are there common factors in individual commodity futures returns? Journal of
Banking & Finance 40 (C), 346–363.
Davis, F. D., Bagozzi, R. P., Warshaw, P. R., 1989. User acceptance of computer technology: A comparison of two theoretical models.
Management Science 35 (8), 982–1003.
Dawid, A. P., 1984. Statistical theory: The prequential approach (with discussion and rejoinder). Journal of the Royal Statistical

190
 Society, Series A 147, 278–292.
Dawid, A. P., DeGroot, M. H., Mortera, J., Cooke, R., French, S., Genest, C., Schervish, M. J., Lindley, D. V., McConway, K. J., Winkler,
R. L., 1995. Coherent combination of experts’ opinions. Test 4 (2), 263–313.
de Albuquerquemello, V. P., de Medeiros, R. K., da Nóbrega Besarria, C., Maia, S. F., 2018. Forecasting crude oil price: Does exist an
optimal econometric model? Energy 155, 578–591.
de Almeida Marques-Toledo, C., Degener, C. M., Vinhal, L., Coelho, G., Meira, W., Codeço, C. T., Teixeira, M. M., 2017. Dengue
prediction by the web: Tweets are a useful tool for estimating and forecasting dengue at country and city level. PLoS Neglected
Tropical Diseases 11 (7), e0005729.
de Almeida Pereira, G. A., Veiga, Á., 2019. Periodic copula autoregressive model designed to multivariate streamflow time series
modelling. Water Resources Management 33 (10), 3417–3431.
De Baets, S., 2019. Surveying forecasting: a review and directions for future research. International Journal of Information and
Decision Sciences.
De Baets, S., Harvey, N., 2020. Using judgment to select and adjust forecasts from statistical models. European Journal of Operational
Research 284 (3), 882–895.
de Brito, M. P., van der Laan, E. A., 2009. Inventory control with product returns: The impact of imperfect information. European
Journal of Operational Research 194 (1), 85–101.
De Gooijer, J., 1998. On threshold moving-average models. Journal of Time Series Analysis 19 (1), 1–18.
De Gooijer, J. G., Hyndman, R. J., 2006. 25 years of time series forecasting. International Journal of Forecasting 22, 443–473.
de Kok, S., 2017. The quest for a better forecast error metric: Measuring more than the average error. Foresight: The International
Journal of Applied Forecasting 46, 36–45.
De Livera, A. M., Hyndman, R. J., Snyder, R. D., 2011. Forecasting time series with complex seasonal patterns using exponential
smoothing. Journal of the American Statistical Association 106 (496), 1513–1527.
De Mare, J., 1980. Optimal prediction of catastrophes with applications to Gaussian processes. Annals of Probability 8 (4), 841–850.
De Menezes, L. M., Bunn, D. W., Taylor, J. W., 2000. Review of guidelines for the use of combined forecasts. European Journal of
Operational Research 120 (1), 190–204.
de Nicola, F., De Pace, P., Hernandez, M. A., 2016. Co-movement of major energy, agricultural, and food commodity price returns: A
time-series assessment. Energy Economics 57 (C), 28–41.
de Oliveira, E. M., Cyrino Oliveira, F. L., 2018. Forecasting mid-long term electric energy consumption through bagging ARIMA and
exponential smoothing methods. Energy 144, 776–788.
de Queiroz, A. R., 2016. Stochastic hydro-thermal scheduling optimization: An overview. Renewable and Sustainable Energy Reviews
62, 382–395.
Dean, J., Ghemawat, S., 2008. MapReduce: Simplified data processing on large clusters. Communications of the ACM 51 (1), 107–113.
Deb, C., Zhang, F., Yang, J., Lee, S. E., Shah, K. W., 2017. A review on time series forecasting techniques for building energy consump-
tion. Renewable and Sustainable Energy Reviews 74, 902–924.
Debecker, A., Modis, T., 1994. Determination of the uncertainties in S-curve logistic fits. Technological Forecasting and Social Change
46 (2), 153–173.
Debecker, A., Modis, T., 2020. Poorly known aspects of flattening the curve of COVID 19. Working Paper.
Degiannakis, S. A., Filis, G., Klein, T., Walther, T., 2020. Forecasting realized volatility of agricultural commodities. International
Journal of Forecasting.
DeGroot, M. H., 2004. Optimal statistical decisions. Wiley-Interscience, Hoboken, N.J.
Dekker, M., van Donselaar, K., Ouwehand, P., 2004. How to use aggregation and combined forecasting to improve seasonal demand
forecasts. International Journal of Production Economics 90 (2), 151–167.
Del Negro, M., Schorfheide, F., 2004. Priors from general equilibrium models for VARS. International Economic Review 45 (2), 643–
673.
Del Negro, M., Schorfheide, F., 2006. How good is what you’ve got? DGSE-VAR as a toolkit for evaluating DSGE models. Economic
Review-Federal Reserve Bank of Atlanta 91 (2), 21.
Del Negro, M., Schorfheide, F., 2013. DSGE model-based forecasting. In: Elliott, G., Timmermann, A. (Eds.), Handbook of Economic
Forecasting, Volume 2. Amsterdam, Horth-Holland, pp. 57–140.
Dellaportas, P., Denison, D. G. T., Holmes, C., 2007. Flexible threshold models for modelling interest rate volatility. Econometric
Reviews 26 (2-4), 419–437.
Delle Monache, L., Hacker, J. P., Zhou, Y., Deng, X., Stull, R. B., 2006. Probabilistic aspects of meteorological and ozone regional
ensemble forecasts. Journal of Geophysical Research: Atmospheres 111 (D24).

191
Demirovic, E., Stuckey, P. J., Bailey, J., Chan, J., Leckie, C., Ramamohanarao, K., Guns, T., 2019. Predict+optimise with ranking
objectives: Exhaustively learning linear functions. In: IJCAI. pp. 1078–1085.
Dempster, A. P., Laird, N. M., Rubin, D. B., 1977. Maximum likelihood from incomplete data via the EM algorithm. Journal of the
Royal Statistical Society, Series B (Statistical Methodology) 39, 1–38.
Di Corso, E., Cerquitelli, T., Apiletti, D., 2018. Metatech: Meteorological data analysis for thermal energy characterization by means
of self-learning transparent models. Energies 11 (6), 1336.
Diab, D. L., Pui, S.-Y., Yankelevich, M., Highhouse, S., 2011. Lay perceptions of selection decision aids in US and Non-US samples.
International Journal of Selection and Assessment 19 (2), 209–216.
Dichtl, H., Drobetz, W., Lohre, H., Rother, C., Vosskamp, P., 2019. Optimal timing and tilting of equity factors. Financial Analysts
Journal 75 (4), 84–102.
Dickey, D. A., Fuller, W. A., 1979. Distribution of the estimators for autoregressive time series with a unit root. Journal of the American
Statistical Association 74 (366), 427–431.
Dickey, D. A., Pantula, S. G., 1987. Determining the order of differencing in autoregressive processes. Journal of Business & Economic
Statistics 5 (4), 455–461.
Diebold, F. X., 2015. Comparing predictive accuracy, twenty years later: A personal perspective on the use and abuse of diebold–
mariano tests. Journal of Business & Economic Statistics 33 (1), 1–1.
Diebold, F. X., Gunther, T. A., Tay, A. S., 1998. Evaluating density forecasts with applications to financial risk management. Interna-
tional Economic Review 39 (4), 863–883.
Diebold, F. X., Mariano, R. S., 1995. Comparing predictive accuracy. Journal of Business & Economic Statistics 13 (3), 253–263.
Diebold, F. X., Pauly, P., 1987. Structural change and the combination of forecasts. Journal of Forecasting 6 (1), 21–40.
Diebold, F. X., Pauly, P., 1990. The use of prior information in forecast combination. International Journal of Forecasting 6 (4), 503–
508.
Diebold, F. X., Shin, M., 2019. Machine learning for regularized survey forecast combination: Partially-egalitarian lasso and its
derivatives. International Journal of Forecasting 35 (4), 1679–1691.
Dieckmann, N. F., Gregory, R., Peters, E., Hartman, R., 2017. Seeing what you want to see: How imprecise uncertainty ranges enhance
motivated reasoning. Risk Analysis 37 (3), 471–486.
Dietrich, J. K., Joines, D. H., 1983. Rational Expectations, Informational Efficiency, and Tests Using Survey Data: A Comment. The
Review of Economics and Statistics 65 (3), 525–529.
Dietvorst, B. J., Simmons, J. P., Massey, C., 2015. Algorithm aversion: people erroneously avoid algorithms after seeing them err.
Journal of Experimental Psychology: General 144 (1), 114–126.
Dietvorst, B. J., Simmons, J. P., Massey, C., 2018. Overcoming algorithm aversion: People will use imperfect algorithms if they can
even slightly modify them. Management Science 64 (3), 1155–1170.
Dietzel, M., Baltzer, P. A., Vag, T., Gröschel, T., Gajda, M., Camara, O., Kaiser, W. A., 2010. Application of breast MRI for prediction
of lymph node metastases–systematic approach using 17 individual descriptors and a dedicated decision tree. Acta Radiologica
51 (8), 885–894.
Ding, A., Hwang, J., 1999. Prediction intervals, factor analysis models, and high-dimensional empirical linear prediction. Journal of
the American Statistical Association 94, 446–455.
Ding, R., Wang, Q., Dang, Y., Fu, Q., Zhang, H., Zhang, D., 2015. Yading: Fast clustering of large-scale time series data. Proceedings
of the VLDB Endowment 8 (5), 473–484.
Dissanayake, G. S., Peiris, M. S., Proietti, T., 2018. Fractionally differenced Gegenbauer processes with long memory: A review.
Statistical Science 33, 413–426.
Divakar, S., Ratchford, B. T., Shankar, V., 2005. CHAN4CAST: A multichannel, multiregion sales forecasting model and decision
support system for consumer packaged goods. Marketing Science 24 (3), 334–350.
Do, L., Vu, H., Vo, B., Liu, Z., Phung, D., 2019. An effective spatial-temporal attention based neural network for traffic flow prediction.
Transportation Research Part C: Emerging technologies 108, 12–28.
Doan, T., Litterman, R., Sims, C., 1984. Forecasting and conditional projection using realistic prior distributions. Econometric Re-
views 3 (1), 1–100.
Dokumentov, A., 2017. Smoothing, decomposition and forecasting of multidimensional and functional time series using regularisa-
tion. Monash University.
Dokumentov, A., Hyndman, R. J., 2018. stR: STR Decomposition. R package version 0.4.
Dolara, A., Grimaccia, F., Leva, S., Mussetta, M., Ogliari, E., 2015. A physical hybrid artificial neural network for short term forecast-
ing of PV plant power output. Energies 8 (2), 1–16.

192
Dolara, A., Grimaccia, F., Leva, S., Mussetta, M., Ogliari, E., 2018. Comparison of training approaches for photovoltaic forecasts by
means of machine learning. Applied Sciences 8 (2), 228.
Dolgin, E., 2010. Better forecasting urged to avoid drug waste.
Dong, X., Li, Y., Rapach, D. E., Zhou, G., 2020. Anomalies and the expected market return.
Doornik, J. A., 2018. Autometrics. In: Castle, J. L., Shephard, N. (Eds.), The methodology and practice of econometrics. Oxford
University Press, Oxford, pp. 88–121.
Doornik, J. A., Castle, J. L., Hendry, D. F., 2020a. Card forecasts for M4. International Journal of Forecasting 36, 129–134.
Doornik, J. A., Castle, J. L., Hendry, D. F., 2020b. Short-term forecasting of the coronavirus pandemic. International Journal of
Forecasting.
Doornik, J. A., Hendry, D. F., 2015. Statistical model selection with “Big Data”. Cogent Economics & Finance 3 (1).
Doucet, A. N. d. F., Gordon, N. J., 2001. Sequential Monte Carlo Methods in Practice. New York: Springer Verlag.
Dowd, K., Cairns, A. J. G., Blake, D., Coughlan, G. D., Epstein, D., Khalaf-Allah, M., 2010. Evaluating the goodness of fit of stochastic
mortality model. Insurance Mathematics and Economics 47 (3), 255–265.
Draper, D., Krnjajić, M., 2013. Calibration results for bayesian model specification. Tech. rep., Department of Applied Mathematics
and Statistics, University of California.
Dudek, G., 2013. Forecasting time series with multiple seasonal cycles using neural networks with local learning. In: International
Conference on Artificial Intelligence and Soft Computing. Springer, pp. 52–63.
Dudek, G., 2015. Generalized regression neural network for forecasting time series with multiple seasonal cycles. In: Intelligent
Systems’2014. Springer, pp. 839–846.
Dudek, G., 2016. Multilayer perceptron for GEFCom2014 probabilistic electricity price forecasting. International Journal of Forecast-
ing 32 (3), 1057–1060.
Duncan, O. D., Davis, B., 1953. An alternative to ecological correlation. American Sociological Review 18, 665–666.
Dunis, C. L., Laws, J., Sermpinis, G., 2010. Modelling and trading the EUR/USD exchange rate at the ECB fixing. The European
Journal of Finance 16 (6), 541–560.
Dunn, D. M., Williams, W. H., Dechaine, T. L., 1976. Aggregate versus subaggregate models in local area forecasting. Journal of the
American Statistical Association 71 (353), 68–71.
Durante, F., Sempi, C., 2015. Principles of copula theory. CRC press.
Durbin, J., Koopman, S. J., 2012. Time series analysis by state space methods. Oxford University Press, Oxford.
e Silva, E. G. d. S., Legey, L. F., e Silva, E. A. d. S., 2010. Forecasting oil price trends using wavelets and hidden Markov models. Energy
Economics 32 (6), 1507–1519.
Eastwood, J., Snook, B., Luther, K., 2012. What people want from their professionals: Attitudes toward decision-making strategies.
Journal of Behavioral Decision Making 25 (5), 458–468.
Eaves, A. H. C., Kingsman, B. G., 2004. Forecasting for the ordering and stock-holding of spare parts. Journal of the Operational
Research Society 55 (4), 431–437.
Economou, T., Stephenson, D. B., Rougier, J. C., Neal, R. A., Mylne, K. R., 2016. On the use of Bayesian decision theory for issu-
ing natural hazard warnings. Proceedings of the Royal Society: Mathematical, Physical, and Engineering Sciences 472 (2194),
20160295.
Edge, R. M., Gürkaynak, R., 2010. How useful are estimated DSGE model forecasts for central bankers? Brookings Papers on Eco-
nomic Activity 41 (2 (Fall)), 209–259.
Edwards, D. G., Hsu, J. C., 1983. Multiple comparisons with the best treatment. Journal of the American Statistical Association
78 (384), 965–971.
Efron, B., 1979. Bootstrap methods: Another look at the jackknife. Annals of Statistics 7 (1), 1–26.
Efron, B., Tibshirani, R., 1986. Bootstrap methods for standard errors, confidence intervals, and other measures of statistical accuracy.
Statistical Science, 54–75.
Eggleton, I. R. C., 1982. Intuitive Time-Series extrapolation. Journal of Accounting Research 20 (1), 68–102.
Ehsani, S., Linnainmaa, J. T., 2020. Factor momentum and the momentum factor. SSRN:3014521.
Eichenbaum, M., Johannsen, B. K., Rebelo, S., 2017. Monetary policy and the predictability of nominal exchange rates. NBER Working
Papers 23158, National Bureau of Economic Research, Inc.
Eksoz, C., Mansouri, S. A., Bourlakis, M., Önkal, D., 2019. Judgmental adjustments through supply integration for strategic partner-
ships in food chains. Omega 87, 20–33.
El Balghiti, O., Elmachtoub, A. N., Grigas, P., Tewari, A., 2019. Generalization bounds in the predict-then-optimize framework. In:
Advances in Neural Information Processing Systems. pp. 14412–14421.

193
El-Hendawi, M., Wang, Z., 2020. An ensemble method of full wavelet packet transform and neural network for short term electrical
load forecasting. Electric Power Systems Research 182, 106265.
Elangasinghe, M. A., Singhal, N., Dirks, K. N., Salmond, J. A., Samarasinghe, S., 2014. Complex time series analysis of PM10 and
PM2.5 for a coastal site using artificial neural network modelling and k-means clustering. Atmospheric Environment 94, 106–116.
Elliott, G., 2015. Complete subset regressions with large-dimensional sets of predictors. Journal of Economic Dynamics & Control
54, 86–111.
Elliott, G., Timmermann, A., Komunjer, I., 2005. Estimation and testing of forecast rationality under flexible loss. The Review of
Economic Studies 72 (4), 1107–1125.
Elliott, M. R., Valliant, R., 2017. Inference for nonprobability samples. Statistical Science 32 (2), 249–264.
Elmachtoub, A. N., Grigas, P., 2017. Smart “predict, then optimize”. arXiv:1710.08005.
Elsbach, K. D., Elofson, G., 2000. How the packaging of decision explanations affects perceptions of trustworthiness. Academy of
Management Journal 43 (1), 80–89.
Embrechts, P., Klüppelberg, C., Mikosch, T., 2013. Modelling Extremal Events: for Insurance and Finance. Springer Science & Busi-
ness Media.
Enders, W., Granger, C. W. J., 1998. Unit-Root tests and asymmetric adjustment with an example using the term structure of interest
rates. Journal of Business & Economic Statistics 16 (3), 304–311.
Engel, C., Lee, D., Liu, C., Liu, C., Wu, S. P. Y., 2019. The uncovered interest parity puzzle, exchange rate forecasting, and Taylor
rules. Journal of International Money and Finance 95, 317–331.
Engel, C., Mark, N. C., West, K. D., 2008. Exchange rate models are not as bad as you think. In: Acemoglu, D., Rogoff, K., Woodford, M.
(Eds.), NBER Macroeconomics Annual 2007. Vol. 22 of NBER Chapters. National Bureau of Economic Research, Inc, pp. 381–441.
Engelberg, J., Manski, C. F., Williams, J., 2009. Comparing the point predictions and subjective probability distributions of profes-
sional forecasters. Journal of Business & Economic Statistics 27 (1), 30–41.
Engle, R., 2002. Dynamic Conditional Correlation. Journal of Business & Economic Statistics 20 (3), 339–350.
Engle, R., 2004. Risk and volatility: Econometric models and financial practice. American Economic Review 94 (3), 405–420.
Engle, R. F., 1982. Autoregressive Conditional Heteroscedasticity with Estimates of the Variance of United Kingdom Inflation. Econo-
metrica 50 (4), 987.
Engle, R. F., Ghysels, E., Sohn, B., 2013. Stock market volatility and macroeconomic fundamentals. Review of Economics and Statistics
95 (3), 776–797.
Engle, R. F., Kroner, K. F. ., 1995. Multivariate Simultaneous Generalized ARCH. Econometric Theory 11 (1), 122–150.
Engle, R. F., Russell, J. R., 1997. Forecasting the frequency of changes in quoted foreign exchange prices with the autoregressive
conditional duration model. Journal of Empirical Finance 4 (2), 187–212.
Engle, R. F., Russell, J. R., 1998. Autoregressive conditional duration: A new model for irregularly spaced transaction data. Econo-
metrica 66 (5), 1127–1162.
Erikson, R. S., Wlezien, C., 2012. Markets vs. polls as election predictors: An historical assessment. Electoral Studies 31 (3), 532–539.
European Banking Federation, 2019. EBF position paper on AI in the banking industry. EBF 037419.
Fahimnia, B., Sanders, N., Siemsen, E., 2020. Human judgment in supply chain forecasting. Omega 94, 102249.
Fair, R. C., 1978. The effect of economic events on votes for president. The Review of Economics and Statistics 60 (2), 159–173.
Fan, S., Chen, L., Lee, W.-J., 2008. Machine learning based switching model for electricity load forecasting. Energy Conversion &
Management 49 (6), 1331–1344.
Fan, S., Mao, C., Chen, L., 2006. Electricity peak load forecasting with self-organizing map and support vector regression. IEEJ
Transactions on Electrical and Electronic Engineering 1 (3), xxxi–xxxi.
Fan, Y., Nowaczyk, S., Röognvaldsson, T., 2020. Transfer learning for remaining useful life prediction based on consensus self-
organizing models. Reliability Engineering and System Safety In Press.
Fan, Y., Tang, C. Y., 2013. Tuning parameter selection in high dimensional penalized likelihood. Journal of the Royal Statistical
Society, Series B (Statistical Methodology) 75 (3), 531–552.
Faraji, J., Ketabi, A., Hashemi-Dezaki, H., Shafie-Khah, M., Catalão, J. P. S., 2020. Optimal Day-Ahead Self-Scheduling and operation
of prosumer microgrids using hybrid machine Learning-Based weather and load forecasting. IEEE Access 8, 157284–157305.
Faust, J., Wright, J. H., 2013. Forecasting inflation. In: Elliott, G., Timmermann, A. (Eds.), Handbook of Economic Forecasting. Vol. 2.
Elsevier, Ch. 1, pp. 2–56.
Favero, C. A., Marcellino, M., 2005. Modelling and forecasting fiscal variables for the euro area. Oxford Bulletin of Economics and
Statistics 67, 755–783.
Fernandes, M., de Sá Mota, B., Rocha, G., 2005. A multivariate conditional autoregressive range model. Economics Letters 86 (3),

194
 435–440.
Fernández-Villaverde, J., Guerrón-Quintana, P. A., 2020. Estimating DSGE models: Recent advances and future challenges. Working
Paper 27715, National Bureau of Economic Research.
Fezzi, C., Mosetti, L., 2020. Size matters: Estimation sample length and electricity price forecasting accuracy. The Energy Journal
41 (4).
Fifić, M., Gigerenzer, G., 2014. Are two interviewers better than one? Journal of Business Research 67 (8), 1771–1779.
Figlewski, S., Wachtel, P., 1981. The Formation of Inflationary Expectations. The Review of Economics and Statistics 63 (1), 1–10.
Figlewski, S., Wachtel, P., 1983. Rational Expectations, Informational Efficiency, and Tests Using Survey Data: A Reply. The Review
of Economics and Statistics 65 (3), 529–531.
Fildes, R., 2017. Research into forecasting practice. Foresight: The International Journal of Applied Forecasting 44, 39–46.
Fildes, R., Goodwin, P., 2007. Against your better judgment? How organizations can improve their use of management judgment in
forecasting. Interfaces 37 (6), 570–576.
Fildes, R., Goodwin, P., 2013. Forecasting support systems: What we know, what we need to know. International Journal of Forecast-
ing 29 (2), 290–294.
Fildes, R., Goodwin, P., Lawrence, M., 2006. The design features of forecasting support systems and their effectiveness. Decision
Support Systems 42 (1), 351–361.
Fildes, R., Goodwin, P., Lawrence, M., Nikolopoulos, K., 2009. Effective forecasting and judgmental adjustments: an empirical evalu-
ation and strategies for improvement in supply-chain planning. International Journal of Forecasting 25 (1), 3–23.
Fildes, R., Goodwin, P., Önkal, D., 2019a. Use and misuse of information in supply chain forecasting of promotion effects. Interna-
tional Journal of Forecasting 35 (1), 144–156.
Fildes, R., Ma, S., Kolassa, S., 2019b. Retail forecasting: research and practice. International Journal of Forecasting.
Fildes, R., Petropoulos, F., 2015. Improving forecast quality in practice. Foresight: The International Journal of Applied Forecasting
36 (Winter), 5–12.
Filippou, I., Rapach, D. E., Taylor, M. P., Zhou, G., 2020. Exchange rate prediction with machine learning and a smart carry trade
portfolio. SSRN:3455713.
Findley, D. F., 2005. Some recent developments and directions in seasonal adjustment. Journal of Official Statistics 21 (2), 343.
Findley, D. F., Monsell, B. C., Bell, W. R., Otto, M. C., Chen, B.-C., 1998. New capabilities and methods of the X-12-ARIMA seasonal-
adjustment program. Journal of Business & Economic Statistics 16 (2), 127–152.
Fiori, C., Kovaka, M., 2005. Defining megaprojects: Learning from construction at the edge of experience. Construction Research
Congress 2005, 1–10.
Fiorucci, J. A., Pellegrini, T. R., Louzada, F., Petropoulos, F., Koehler, A. B., 2016. Models for optimising the theta method and their
relationship to state space models. International Journal of Forecasting 32 (4), 1151–1161.
Fioruci, J. A., Pellegrini, T. R., Louzada, F., Petropoulos, F., 2015. The optimised theta method. arXiv:1503.03529.
Firebaugh, G., 1978. A rule for inferring Individual-Level relationships from aggregate data. American Sociological Review 43 (4),
557–572.
Fischhoff, B., 2007. An early history of hindsight research. Social Cognition 25 (1), 10–13.
Fischhoff, B., 2012. Communicating uncertainty fulfilling the duty to inform. Issues in Science and Technology 28 (4), 63–70.
Fischhoff, B., Davis, A. L., 2014. Communicating scientific uncertainty. Proceedings of the National Academy of Sciences 111 (Sup-
plement 4), 13664–13671.
Fisher, J. C., Pry, R. H., 1971. A simple substitution model of technological change. Technological Forecasting and Social Change 3,
75–88.
Fiske, S. T., Dupree, C., 2014. Gaining trust as well as respect in communicating to motivated audiences about science topics. Pro-
ceedings of the National Academy of Sciences 111 (Supplement 4), 13593–13597.
Fiszeder, P., 2005. Forecasting the volatility of the Polish stock index – WIG20. In: Milo, W., Wdowiński, P. (Eds.), Forecasting
Financial Markets. Theory and Applications. Wydawnictwo Uniwersytetu Łódzkiego, pp. 29–42.
Fiszeder, P., 2018. Low and high prices can improve covariance forecasts: The evidence based on currency rates. Journal of Forecasting
37 (6), 641–649.
Fiszeder, P., Fałdziński, M., 2019. Improving forecasts with the co-range dynamic conditional correlation model. Journal of Economic
Dynamics and Control 108, 103736.
Fiszeder, P., Fałdziński, M., Molnár, P., 2019. Range-based DCC models for covariance and value-at-risk forecasting. Journal of Em-
pirical Finance 54, 58–76.
Fiszeder, P., Perczak, G., 2013. A new look at variance estimation based on low, high and closing prices taking into account the drift.

195
 Statistica Neerlandica 67 (4), 456–481.
Fiszeder, P., Perczak, G., 2016. Low and high prices can improve volatility forecasts during periods of turmoil. International Journal
of Forecasting 32 (2), 398–410.
Fixler, D. J., Grimm, B. T., 2005. Reliability of the NIPA estimates of U.S. economic activity. Survey of Current Business 85, 9–19.
Fixler, D. J., Grimm, B. T., 2008. The reliability of the GDP and GDI estimates. Survey of Current Business 88, 16–32.
Fliedner, G., 2003. CPFR: an emerging supply chain tool. Industrial Management & Data Systems 103 (1), 14–21.
Floros, C., 2005. Forecasting the UK unemployment rate: Model comparisons. Journal of Applied Econometrics and Quantitative
Studies 2-4, 57–72.
Flyvbjerg, B., 2007. Policy and planning for Large-Infrastructure projects: Problems, causes, cures. Environment and Planning: B,
Planning & Design 34 (4), 578–597.
Flyvbjerg, B., Bruzelius, N., Rothengatter, W., 2003. Megaprojects and Risk: An Anatomy of Ambition. Cambridge University Press.
Forcina, A., Pellegrino, D., 2019. Estimation of voter transitions and the ecological fallacy. Quality & Quantity 53 (4), 1859–1874.
Forni, M., Hallin, M., Lippi, M., Reichlin, L., 2000. The generalized factor model: Identification and estimation. The Review of
Economics and Statistics 82, 540–554.
Forrest, D., Goddard, J., Simmons, R., 2005. Odds-Setters As Forecasters: The Case of English Football. International Journal of
Forecasting 21, 551–564.
Fortsch, S. M., Khapalova, E. A., 2016. Reducing uncertainty in demand for blood. Operations Research for Health Care 9, 16–28.
Fortuin, L., 1984. Initial supply and re-order level of new service parts. European Journal of Operational Research 15 (3), 310–319.
Foucquier, A., Robert, S., Suard, F., Stéphan, L., Jay, A., 2013. State of the art in building modelling and energy performances predic-
tion: A review. Renewable and Sustainable Energy Reviews 23, 272–288.
Fox, A. J., 1972. Outliers in time series. Journal of the Royal Statistical Society, Series B (Statistical Methodology) 34 (3), 350–363.
Frankel, J., Schreger, J., 2013. Over-optimistic official forecasts and fiscal rules in the eurozone. Review of World Economics 149,
247–272.
Franses, P. H., 1991. Seasonality, non-stationarity and the forecasting of monthly time series. International Journal of Forecasting
7 (2), 199–208.
Franses, P. H., 1994. A multivariate approach to modeling univariate seasonal time series. Journal of Econometrics 63 (1), 133–151.
Franses, P. H., Ghijsels, H., 1999. Additive outliers, GARCH and forecasting volatility. International Journal of Forecasting 15 (1),
1–9.
Franses, P. H., Hoek, H., Paap, R., 1997. Bayesian analysis of seasonal unit roots and seasonal mean shifts. Journal of Econometrics
78 (2), 359–380.
Franses, P. H., Kleibergen, F., 1996. Unit roots in the Nelson-Plosser data: Do they matter for forecasting? International Journal of
Forecasting 12 (2), 283–288.
Franses, P. H., Legerstee, R., 2009a. Do experts’ adjustments on model-based SKU-level forecasts improve forecast quality? Journal
of Forecasting 36.
Franses, P. H., Legerstee, R., 2009b. Properties of expert adjustments on model-based SKU-level forecasts. International Journal of
Forecasting 25 (1), 35–47.
Franses, P. H., van Dijk, D., Opschoor, A., 2014. Time Series Models for Business and Economic Forecasting. Cambridge University
Press.
Franses, P. H., Vogelsang, T. J., 1998. On seasonal cycles, unit roots, and mean shifts. The Review of Economics and Statistics 80 (2),
231–240.
Frechtling, D. C., 2001. Forecasting Tourism Demand: Methods and Strategies. Routledge.
Freedman, D. A., 1981. Bootstrapping regression models. The Annals of Statistics 9 (6), 1218–1228.
Freedman, D. A., Klein, S. P., Ostland, M., Roberts, M., 1998. Review of ‘A Solution to the Ecological Inference Problem’. Journal of
the American Statistical Association 93 (444), 1518–1522.
Freyberger, J., Neuhierl, A., Weber, M., 2020. Dissecting characteristics nonparametrically. Review of Financial Studies 33, 2326–
2377.
Friedman, J. A., 2015. Using power laws to estimate conflict size. The Journal of Conflict Resolution 59 (7), 1216–1241.
Fry, C., Brundage, M., 2020. The M4 forecasting competition – A practitioner’s view. International Journal of Forecasting 36 (1),
156–160.
Fuhrer, J. C., 2018. Intrinsic Expectations Persistence: Evidence from Professional and Household Survey Expectations. Working
Papers 18-9, Federal Reserve Bank of Boston.
Fulcher, B. D., Jones, N. S., 2014. Highly comparative feature-based time-series classification. IEEE Transactions on Knowledge and

196
 Data Engineering 26 (12), 3026–3037.
Fulcher, B. D., Little, M. A., Jones, N. S., 2013. Highly comparative time-series analysis: the empirical structure of time series and
their methods. Journal of the Royal Society Interface 10 (83), 20130048.
Funahashi, K.-I., 1989. On the approximate realization of continuous mappings by neural networks. Neural Networks 2 (3), 183–192.
Furlan, C., Guidolin, M., Guseo, R., 2016. Has the Fukushima accident influenced short-term consumption in the evolution of nuclear
energy an analysis of the world and seven leading countries. Technological Forecasting and Social Change 107, 37–49.
Furlan, C., Mortarino, C., 2018. Forecasting the impact of renewable energies in competition with non-renewable sources. Renewable
and Sustainable Energy Reviews 81, 1879–1886.
Furlan, C., Mortarino, C., Zahangir, M., 2020. Interaction among three substitute products: An extended innovation diffusion model.
Statistical Methods and Applications in press.
Gaddis, J. L., 1989. The Long Peace: Inquiries Into the History of the Cold War. The Long Peace: Inquiries Into the History of the Cold
War.
Gaillard, P., Goude, Y., Nedellec, R., 2016. Additive models and robust aggregation for GEFCom2014 probabilistic electric load and
electricity price forecasting. International Journal of forecasting 32 (3), 1038–1050.
Galbreth, M. R., Kurtuluş, M., Shor, M., 2015. How collaborative forecasting can reduce forecast accuracy. Operations Research Letters
43 (4), 349–353.
Gali, J., 2008. Monetary Policy, Inflation, and the Business Cycle: An Introduction to the New Keynesian Framework. Princeton
University Press, Princeton and Oxford.
Galicia, A., Talavera-Llames, R., Troncoso, A., Koprinska, I., Martı́nez-Álvarez, F., 2019. Multi-step forecasting for big data time series
based on ensemble learning. Knowledge-Based Systems 163, 830–841.
Galicia, A., Torres, J. F., Martı́nez-Álvarez, F., Troncoso, A., 2018. A novel Spark-based multi-step forecasting algorithm for big data
time series. Information Sciences 467, 800–818.
Galvão, A. B., 2017. Data revisions and DSGE models. Journal of Econometrics 196 (1), 215–232.
Galvão, A. B., Giraitis, L., Kapetanios, G., Petrova, K., 2016. A time varying DSGE model with financial frictions. Journal of Empirical
Finance 38, 690–716.
Gamble, C., Gao, J., 2018. Safety-first AI for autonomous data centre cooling and industrial control. https://deepmind.com/blog/
article/safety-first-ai-autonomous-data-centre-cooling-and-industrial-control, accessed on 2020-09-01.
Gans, N., Koole, G., Mandelbaum, A., 2003. Telephone call centers: Tutorial, review, and research prospects. Manufacturing & Service
Operations Management 5 (2), 79–141.
Garcı́a, F. P., Pedregal, D. J., Roberts, C., 2010. Time series methods applied to failure prediction and detection. Reliability Engineering
& System Safety 95 (6), 698–703.
Garcia, R., Perron, P., 1996. An analysis of the real interest rate under regime shifts. The Review of Economics and Statistics 78 (1),
111–125.
Gardner, Jr, E., Koehler, A. B., 2005. Comments on a patented bootstrapping method for forecasting intermittent demand. Interna-
tional Journal of Forecasting 21 (3), 617–618.
Gardner, E. S., 1985. Exponential smoothing: The state of the art. Journal of Forecasting 4 (1), 1–28.
Gardner, E. S., 2006. Exponential smoothing: The state of the art - part II. International Journal of Forecasting 22 (4), 637–666.
Garman, M. B., Klass, M. J., 1980. On the estimation of security price volatilities from historical data. The Journal of Business 53 (1),
67–78.
Garratt, A., Lee, K., Mise, E., Shields, K., 2008. Real time representations of the output gap. Review of Economics and Statistics 90,
792–804.
Gasthaus, J., Benidis, K., Wang, Y., Rangapuram, S. S., Salinas, D., Flunkert, V., Januschowski, T., 2019. Probabilistic forecasting with
spline quantile function RNNs. In: The 22nd International Conference on Artificial Intelligence and Statistics. pp. 1901–1910.
Gebicki, M., Mooney, E., Chen, S.-J. G., Mazur, L. M., 2014. Evaluation of hospital medication inventory policies. Health Care Man-
agement Science 17 (3), 215–229.
Gelman, A., Park, D. K., Ansolabehere, S., Price, P. N., Minnite, L. C., 2001. Models, assumptions and model checking in ecological
regressions. Journal of the Royal Statistical Society, Series A, 164 (1), 101–118.
Gelper, S., Fried, R., Croux, C., 2009. Robust forecasting with exponential and Holt-Winters smoothing. Journal of Forecasting 11.
Gentine, P., Pritchard, M., Rasp, S., Reinaudi, G., Yacalis, G., 2018. Could machine learning break the convection parameterization
deadlock? Geophysical Research Letters 45 (11), 5742–5751.
George, E. I., McCulloch, R. E., 1993. Variable selection via Gibbs sampling. Journal of the American Statistical Association 88 (423),
881–890.

197
Gerlach, R., Chen, C. W. S., Lin, D. S. Y., Huang, M.-H., 2006. Asymmetric responses of international stock markets to trading volume.
Physica A: Statistical Mechanics and its Applications 360 (2), 422–444.
Geweke, J., 1977. The dynamic factor analysis of economic time series. Latent variables in socio-economic models.
Geweke, J., 2001. Bayesian econometrics and forecasting. Journal of Econometrics 100 (1), 11–15.
Geweke, J., Amisano, G., 2010. Comparing and evaluating Bayesian predictive distributions of asset returns. International Journal of
Forecasting 26 (2), 216–230.
Geweke, J., Amisano, G., 2011. Optimal prediction pools. Journal of Econometrics 164 (1), 130–141.
Gharbi, M., Quenel, P., Gustave, J., Cassadou, S., La Ruche, G., Girdary, L., Marrama, L., 2011. Time series analysis of dengue
incidence in guadeloupe, french west indies: forecasting models using climate variables as predictors. BMC Infectious Diseases
11 (1), 1–13.
Ghassemi, M., Pimentel, M. A., Naumann, T., Brennan, T., Clifton, D. A., Szolovits, P., Feng, M., 2015. A multivariate timeseries
modeling approach to severity of illness assessment and forecasting in icu with sparse, heterogeneous clinical data. In: Proceedings
of the AAAI Conference on Artificial Intelligence. AAAI Conference on Artificial Intelligence. Vol. 2015. NIH Public Access, p.
446.
Ghysels, E., Lee, H. S., Noh, J., 1994. Testing for unit roots in seasonal time series: Some theoretical extensions and a Monte Carlo
investigation. Journal of Econometrics 62 (2), 415–442.
Ghysels, E., Plazzi, A., Valkanov, R., Torous, W., 2013. Forecasting real estate prices. In: Elliott, G., Timmermann, A. (Eds.), Handbook
of Economic Forecasting. Vol. 2. Elsevier, pp. 509–580.
Giacomini, R., Rossi, B., 2016. Model comparisons in unstable environments. International Economic Review 57 (2), 369–392.
Giacomini, R., White, H., 2006. Tests of conditional predictive ability. Econometrica 74 (6), 1545–1578.
Giannone, D. L., Primiceri, G. M., 2017. Macroeconomic prediction with big data: The illusion of sparsity. The Fedral Reserve Bank
of New York.
Gias, A. U., Casale, G., 2020. Cocoa: Cold start aware capacity planning for function-as-a-service platforms. arXiv:2007.01222.
Gigerenzer, G., 1996. On narrow norms and vague heuristics: A reply to Kahneman and Tversky. Psychological Review 103 (3),
592–596.
Gigerenzer, G., 2007. Gut Feelings: The Intelligence of the Unconscious. Viking.
Gil, R. G. R., Levitt, S. D., 2007. Testing the efficiency of markets in the 2002 World Cup. The Journal of Prediction Markets 1 (3),
255–270.
Gil-Alana, L., 2001. A fractionally integrated exponential model for UK unemployment. Journal of Forecasting 20 (5), 329–340.
Gilbert, C., Browell, J., McMillan, D., 2020a. Leveraging turbine-level data for improved probabilistic wind power forecasting. IEEE
Transactions on Sustainable Energy 11 (3), 1152–1160.
Gilbert, C., Browell, J., McMillan, D., 2020b. Probabilistic access forecasting for improved offshore operations. International Journal
of Forecasting.
Gilbert, K., 2005. An ARIMA supply chain model. Management Science 51 (2), 305–310.
Gilliland, M., 2002. Is forecasting a waste of time? Supply Chain Management Review 6 (4), 16–23.
Gilliland, M., 2010. The Business Forecasting Deal: Exposing Myths, Eliminating Bad Practices, Providing Practical Solutions. John
Wiley & Sons.
Giraitis, L., Kapetanios, G., Price, S., 2013. Adaptive Forecasting in the Presence of Recent and Ongoing Structural Change. Journal
of Econometrics 177 (2), 153–170.
Givon, M., Mahajan, W., Müller, E., 1995. Software piracy: Estimation of the lost sales and the impact on software diffusion. Journal
of Marketing 59, 29–37.
Glahn, H. R., Lowry, D. A., 1972. The use of model output statistics (MOS) in objective weather forecasting. Journal of Applied
Meteorology 11 (8), 1203–1211.
Glasgow, H. B., Burkholder, J. M., Reed, R. E., Lewitus, A. J., Kleinman, J. E., 2004. Real-time remote monitoring of water quality:
a review of current applications, and advancements in sensor, telemetry, and computing technologies. Journal of Experimental
Marine Biology and Ecology 300 (1-2), 409–448.
Glocker, C., Wegmüller, P., 2018. International evidence of time-variation in trend labor productivity growth. Economics Letters 167,
115–119.
Glosten, L. R., Jagannathan, R., Runkle, D. E., 1993. On the relation between the expected value and the volatility of the nominal
excess return on stocks. The Journal of Finance 48 (5), 1779–1801.
Glynn, A., Wakefield, J., 2010. Ecological inference in the social sciences. Statistical Methodology 7 (3), 307–322.
Gneiting, T., 2011a. Making and evaluating point forecasts. Journal of the American Statistical Association 106 (494), 746–762.

198
Gneiting, T., 2011b. Quantiles as optimal point forecasts. International Journal of Forecasting 27 (2), 197–207.
Gneiting, T., Balabdaoui, F., Raftery, A. E., 2007. Probabilistic forecasts, calibration and sharpness. Journal of the Royal Statistical
Society, Series B (Statistical Methodology) 69, 243–268.
Gneiting, T., Katzfuss, M., 2014. Probabilistic forecasting. Annual Review of Statistics and Its Application 1, 125–151.
Gneiting, T., Raftery, A. E., 2007. Strictly proper scoring rules, prediction, and estimation. Journal of the American Statistical Associ-
ation 102 (477), 359–378.
Gneiting, T., Raftery, A. E., Westveld, A. H., Goldman, T., 2005. Calibrated probabilistic forecasting using ensemble model output
statistics and minimum CRPS estimation. Monthly Weather Review 133 (5), 1098–1118.
Gneiting, T., Ranjan, R., 2013. Combining predictive distributions. Electronic Journal of Statistics 7, 1747–1782.
Gneiting, T., Stanberry, L. I., Grimit, E. P., Held, L., Johnson, N. A., 2008. Assessing probabilistic forecasts of multivariate quantities,
with applications to ensemble predictions of surface winds (with discussion and rejoinder). Test 17, 211–264.
Godahewa, R., Deng, C., Prouzeau, A., Bergmeir, C., 2020. Simulation and optimisation of air conditioning systems using machine
learning. arXiv:2006.15296.
Godbole, N., Srinivasaiah, M., Skiena, S., 2007. Large-scale sentiment analysis for news and blogs. ICWSM 7 (21), 219–222.
Godet, M., 1982. From forecasting to ‘la prospective’ a new way of looking at futures. Journal of Forecasting 1 (3), 293–301.
Goh, T. N., Varaprasad, N., 1986. A statistical methodology for the analysis of the Life-Cycle of reusable containers. IIE Transactions
18 (1), 42–47.
Gohin, A., Chantret, F., 2010. The long-run impact of energy prices on world agricultural markets: The role of macro-economic
linkages. Energy Policy 38 (1), 333–339.
Goia, A., May, C., Fusai, G., 2010. Functional clustering and linear regression for peak load forecasting. International Journal of
Forecasting 26 (4), 700–711.
Golan, A., Perloff, J. M., 2004. Superior forecasts of the U.S. unemployment rate using a nonparametric method. The Review of
Economics and Statistics 86 (1), 433–438.
Goldberg, Y., 2017. Neural network methods for natural language processing. Synthesis Lectures on Human Language Technologies
10 (1), 1–309.
Goldstein, D. G., Gigerenzer, G., 2002. Models of ecological rationality: the recognition heuristic. Psychological Review 109 (1),
75–90.
Goldstein, J. S., 2011. Winning the War on War: The Decline of Armed Conflict Worldwide. Penguin.
Golestaneh, F., Pinson, P., Gooi, H. B., 2019. Polyhedral predictive regions for power system applications. IEEE Transactions on Power
Systems 34 (1), 693–704.
Goltsos, T., Syntetos, A., 2020. Forecasting for remanufacturing. Foresight: The International Journal of Applied Forecasting 56,
10–17.
Goltsos, T. E., Syntetos, A. A., van der Laan, E., 2019. Forecasting for remanufacturing: The effects of serialization. Journal of Opera-
tions Management 65 (5), 447–467.
Gomez Munoz, C. Q., De la Hermosa Gonzalez-Carrato, R. R., Trapero Arenas, J. R., Garcia Marquez, F. P., 2014. A novel approach to
fault detection and diagnosis on wind turbines. GlobalNEST International Journal 16 (6), 1029–1037.
Gonçalves, C., Bessa, R. J., Pinson, P., 2020a. A critical overview of privacy-preserving approaches for collaborative forecasting.
International Journal of Forecasting.
Gonçalves, C., Pinson, P., Bessa, R. J., 2020b. Towards data markets in renewable energy forecasting. IEEE Transactions on Sustainable
Energy.
Gönül, M. S., Önkal, D., Goodwin, P., 2009. Expectations, use and judgmental adjustment of external financial and economic forecasts:
an empirical investigation. Journal of Forecasting 28 (1), 19–37.
Gönül, M. S., Önkal, D., Goodwin, P., 2012. Why should I trust your forecasts? Foresight: The International Journal of Applied
Forecasting 27, 5–9.
Gönül, M. S., Önkal, D., Lawrence, M., 2006. The effects of structural characteristics of explanations on use of a DSS. Decision Support
Systems 42 (3), 1481–1493.
Goodman, L. A., 1953. Ecological regressions and behavior of individuals. American Sociological Review 18, 663–664.
Goodman, L. A., 1959. Some alternatives to ecological correlation. The American Journal of Sociology 64 (6), 610–625.
Goodwin, P., 2000a. Correct or combine? mechanically integrating judgmental forecasts with statistical methods. International Jour-
nal of Forecasting 16 (2), 261–275.
Goodwin, P., 2000b. Improving the voluntary integration of statistical forecasts and judgment. International Journal of Forecasting
16 (1), 85–99.

199
Goodwin, P., 2002. Integrating management judgment and statistical methods to improve short-term forecasts. Omega 30 (2), 127–
135.
Goodwin, P., 2014. Getting real about uncertainty. Foresight: The International Journal of Applied Forecasting 33, 4–7.
Goodwin, P., Dyussekeneva, K., Meeran, S., 2013a. The use of analogies in forecasting the annual sales of new electronics products.
IMA Journal of Management Mathematics 24 (4), 407–422.
Goodwin, P., Fildes, R., 1999. Judgmental forecasts of time series affected by special events: does providing a statistical forecast
improve accuracy? Journal of Behavioural Decision Making 12 (1), 37–53.
Goodwin, P., Fildes, R., Lawrence, M., Nikolopoulos, K., 2007. The process of using a forecasting support system. International
Journal of Forecasting 23 (3), 391–404.
Goodwin, P., Fildes, R., Lawrence, M., Stephens, G., 2011. Restrictiveness and guidance in support systems. Omega 39 (3), 242–253.
Goodwin, P., Gönül, M. S., Önkal, D., 2013b. Antecedents and effects of trust in forecasting advice. International Journal of Forecast-
ing 29 (2), 354–366.
Goodwin, P., Gönül, M. S., Önkal, D., 2019a. When providing optimistic and pessimistic scenarios can be detrimental to judgmental
demand forecasts and production decisions. European Journal of Operational Research 273 (3), 992–1004.
Goodwin, P., Gönül, M. S., Önkal, D., Kocabıyıkoğlu, A., Göğüş, I., 2019b. Contrast effects in judgmental forecasting when assessing
the implications of worst- and best-case scenarios. Journal of Behavioral Decision Making 32 (5), 536–549.
Goodwin, P., Meeran, S., Dyussekeneva, K., 2014. The challenges of pre-launch forecasting of adoption time series for new durable
products. International Journal of Forecasting 30 (4), 1082–1097.
Goodwin, P., Petropoulos, F., Hyndman, R. J., 2017. A note on upper bounds for forecast-value-added relative to naı̈ve forecasts.
Journal of the Operational Research Society 68 (9), 1082–1084.
Goodwin, P., Wright, G., 2010. The limits of forecasting methods in anticipating rare events. Technological Forecasting and Social
Change 77 (3), 355–368.
Google code, 2013. The Word2Vec project. https://code.google.com/archive/p/word2vec/, accessed on 2020-09-05.
Gordon, R. J., 2003. Exploding productivity growth: context, causes, and implications. Brookings Papers on Economic Activity
2003 (2), 207–298.
Gospodinov, N., 2005. Testing for threshold nonlinearity in Short-Term interest rates. Journal of Financial Econometrics 3 (3), 344–
371.
Gould, P. G., Koehler, A. B., Ord, J. K., Snyder, R. D., Hyndman, R. J., Vahid-Araghi, F., 2008. Forecasting time series with multiple
seasonal patterns. European Journal of Operational Research 191 (1), 207–222.
Goyal, A., Welch, I., 2008. A comprehensive look at the empirical performance of equity premium prediction. Review of Financial
Studies 21 (4), 1455–1508.
Graefe, A., 2014. Accuracy of vote expectation surveys in forecasting elections. Public Opinion Quarterly 78 (1), 204–232.
Graefe, A., Armstrong, J. S., 2011. Comparing face-to-face meetings, nominal groups, Delphi and prediction markets on an estimation
task. International Journal of Forecasting 27 (1), 183–195.
Graefe, A., Armstrong, J. S., Jones Jr, R. J., Cuzán, A. G., 2014. Combining forecasts: An application to elections. International Journal
of Forecasting 30 (1), 43–54.
Granger, C. W., Ramanathan, R., 1984. Improved methods of combining forecasts. Journal of Forecasting 3 (2), 197–204.
Granger, C. W. J., 1969. Investigating causal relations by econometric models and cross-spectral methods. Econometrica 37 (3), 424–
438.
Granger, C. W. J., Pesaran, M. H., 2000. Economic and statistical measures of forecast accuracy. Journal of Forecasting 19, 537–560.
Granger, C. W. J., Swanson, N., 1996. Future developments in the study of cointegrated variables. Oxford Bulletin of Economics and
Statistics 58 (3), 537–553.
Graves, S. C., 1999. A Single-Item inventory model for a nonstationary demand process. Manufacturing & Service Operations Man-
agement 1 (1), 50–61.
Gray, C. W., Barnes, C. B., Wilkinson, E. F., 1965. The process of prediction as a function of the correlation between two scaled
variables. Psychonomic Science 3 (1), 231–231.
Gray, J., 2015a. Heresies: Against Progress And Other Illusions. Granta Books.
Gray, J., 2015b. Steven Pinker is wrong about violence and war. http://www.theguardian.com/books/2015/mar/13/
john-gray-steven-pinker-wrong-violence-war-declining, accessed on 2018-05-02.
Gray, S. F., 1996. Modeling the conditional distribution of interest rates as a regime-switching process. Journal of Financial Economics
42 (1), 27–62.
Green, J., Hand, J. R. M., Zhang, X. F., 2017. The characteristics that provide independent information about average u.s. monthly

200
 stock returns. Review of Financial Studies 30 (12), 4389–4436.
Green, K. C., Armstrong, J. S., 2007. Structured analogies for forecasting. International Journal of Forecasting 23 (3), 365–376.
Green, K. C., Armstrong, J. S., 2015. Simple versus complex forecasting: The evidence. Journal of Business Research 68 (8), 1678–
1685.
Greiner, D. J., 2007. Ecological inference in voting rights act disputes: Where are we now, and where do we want to be? Jurimetrics
47 (2), 115–167.
Greiner, D. J., Quinn, K. M., 2010. Exit polling and racial bloc voting: combining individual-level and RxC ecological data. The
Annals of Applied statistics 4 (4), 1774–1796.
Gresnigt, F., Kole, E., Franses, P. H., 2015. Interpreting financial market crashes as earthquakes: A new early warning system for
medium term crashes. Journal of Banking & Finance 56, 123–139.
Gresnigt, F., Kole, E., Franses, P. H., 2017a. Exploiting spillovers to forecast crashes. Journal of Forecasting 36 (8), 936–955.
Gresnigt, F., Kole, E., Franses, P. H., 2017b. Specification testing in hawkes models. Journal of Financial Econometrics 15 (1), 139–171.
Gromenko, O., Kokoszka, P., Reimherr, M., 2017. Detection of change in the spatiotemporal mean function. Journal of the Royal
Statistical Society, Series B (Statistical Methodology) 79 (1), 29–50.
Gross, C. W., Sohl, J. E., 1990. Disaggregation methods to expedite product line forecasting. Journal of Forecasting 9 (3), 233–254.
Grossi, L., Nan, F., 2019. Robust forecasting of electricity prices: Simulations, models and the impact of renewable sources. Techno-
logical Forecasting and Social Change 141, 305–318.
Grushka-Cockayne, Y., Jose, V. R. R., 2020. Combining prediction intervals in the M4 competition. International Journal of Forecast-
ing 36 (1), 178–185.
Grushka-Cockayne, Y., Jose, V. R. R., Lichtendahl, K. C., 2017a. Ensembles of overfit and overconfident forecasts. Management Science
63 (4), 1110–1130.
Grushka-Cockayne, Y., Lichtendahl, K. C., Jose, V. R. R., Winkler, R. L., 2017b. Quantile evaluation, sensitivity to bracketing, and
sharing business payoffs. Operations Research 65 (3), 712–728.
Gu, S., Kelly, B., Xiu, D., 2020. Empirical asset pricing via machine learning. Review of Financial Studies 33 (5), 2223–2273.
Guidolin, M., Alpcan, T., 2019. Transition to sustainable energy generation in Australia: Interplay between coal, gas and renewables.
Renewable Energy 139, 359–367.
Guidolin, M., Guseo, R., 2012. A nuclear power renaissance? Technological Forecasting and Social Change 79 (9), 1746–1760.
Guidolin, M., Guseo, R., 2015. Technological change in the U.S. music industry: Within-product, cross–product and churn effects
between competing blockbusters. Technological Forecasting and Social Change 99, 35–46.
Guidolin, M., Guseo, R., 2016. The German energy transition: Modeling competition and substitution between nuclear power and
renewable energy technologies. Renewable and Sustainable Energy Reviews 60, 1498–1504.
Guidolin, M., Guseo, R., 2020. Has the iPhone cannibalized the iPad? An asymmetric competition model. Applied Stochastic Models
in Business and Industry 36, 465–476.
Guidolin, M., Mortarino, C., 2010. Cross-country diffusion of photovoltaic systems: modelling choices and forecasts for national
adoption patterns. Technological Forecasting and Social Change 77 (2), 279–296.
Guidolin, M., Pedio, M., 2018. Essentials of Time Series for Financial Applications. Academic Press.
Guidolin, M., Pedio, M., 2019. Forecasting and trading monetary policy effects on the riskless yield curve with regime switching
Nelson–Siegel models. Journal of Economic Dynamics & Control 107, 103723.
Guidolin, M., Thornton, D. L., 2018. Predictions of short-term rates and the expectations hypothesis. International Journal of Fore-
casting 34 (4), 636–664.
Guidolin, M., Timmermann, A., 2006. Term structure of risk under alternative econometric specifications. Journal of Econometrics
131 (1), 285–308.
Guidolin, M., Timmermann, A., 2009. Forecasts of US short-term interest rates: A flexible forecast combination approach. Journal of
Econometrics 150 (2), 297–311.
Gumus, M., Kiran, M. S., 2017. Crude oil price forecasting using XGBoost. In: 2017 International Conference on Computer Science
and Engineering (UBMK). IEEE, pp. 1100–1103.
Gunter, U., Önder, I., 2016. Forecasting city arrivals with Google Analytics. Annals Of Tourism Research 61, 199–212.
Gunter, U., Önder, I., Gindl, S., 2019. Exploring the predictive ability of LIKES of posts on the facebook pages of four major city
DMOs in Austria. Tourism Economics 25 (3), 375–401.
Guo, X., Grushka-Cockayne, Y., De Reyck, B., 2020. Forecasting airport transfer passenger flow using real-time data and machine
learning. Manufacturing & Service Operations Management.
Gupta, M., Gao, J., Aggarwal, C. C., Han, J., 2013. Outlier detection for temporal data: A survey. IEEE Transactions on Knowledge

201
 and Data Engineering 26 (9), 2250–2267.
Gupta, S., 1994. Managerial judgment and forecast combination: An experimental study. Marketing Letters 5 (1), 5–17.
Gur Ali, O., Pinar, E., 2016. Multi-period-ahead forecasting with residual extrapolation and information sharing — utilizing a multi-
tude of retail series. International Journal of Forecasting 32 (2), 502–517.
Guseo, R., Dalla Valle, A., Guidolin, M., 2007. World oil depletion models: price effects compared with strategic or technological
interventions. Technological Forecasting and Social Change 74 (4), 452–469.
Guseo, R., Guidolin, M., 2009. Modelling a dynamic market potential: A class of automata networks for diffusion of innovations.
Technological Forecasting and Social Change 76, 806–820.
Guseo, R., Guidolin, M., 2011. Market potential dynamics in innovation diffusion: Modelling the synergy between two driving forces.
Technological Forecasting and Social Change 78, 13–24.
Guseo, R., Mortarino, C., 2010. Correction to the paper “Optimal product launch times in a duopoly: Balancing life-cycle revenues
with product cost”. Operations Research 58, 1522–1523.
Guseo, R., Mortarino, C., 2012. Sequential market entries and competition modelling in multi-innovation diffusions. European Jour-
nal of Operational Research 216, 658–667.
Guseo, R., Mortarino, C., 2014. Within-brand and cross-brand word-of-mouth for sequential multi-innovation diffusions. IMA Jour-
nal of Management Mathematics 25, 287–311.
Guseo, R., Mortarino, C., 2015. Modeling competition between two pharmaceutical drugs using innovation diffusion models. The
Annals of Applied Statistics 9, 2073–2089.
Gutierrez, R. S., Solis, A. O., Mukhopadhyay, S., 2008. Lumpy demand forecasting using neural networks. International Journal of
Production Economics 111 (2), 409–420.
Gutterman, S., Vanderhoof, I. T., 1998. Forecasting changes in mortality: A search for a law of causes and effects. North American
Actuarial Journal 2 (4), 135–138.
Gürkaynak, R. S., Kısacıkoğlu, B., Rossi, B., 2013. Do DSGE models forecast more accurately out-of-sample than var models? Ad-
vances in Econometrics,VAR Models in Macroeconomics – New Developments and Applications: Essays in Honor of Christopher
A. Sims 32, 27–79.
Haas, M., Mittnik, S., Paolella, M. S., 2004. A new approach to Markov-Switching GARCH models. Journal of Financial Econometrics
2 (4), 493–530.
Hahn, H., Meyer-Nieberg, S., Pickl, S., 2009. Electric load forecasting methods: Tools for decision making. European Journal of
Operational Research 199 (3), 902–907.
Hahn, M., Frühwirth-Schnatter, S., Sass, J., 2010. Markov chain Monte Carlo methods for parameter estimation in multidimensional
continuous time markov switching models. Journal of Financial Econometrics 8 (1), 88–121.
Hall, P., 1990. Using the bootstrap to estimate mean squared error and select smoothing parameter in nonparametric problems.
Journal of Multivariate Analysis 32 (2), 177–203.
Hall, S. G., Mitchell, J., 2007. Combining density forecasts. International Journal of Forecasting 23 (1), 1–13.
Hall, S. G., Mitchell, J., 2009. Recent developments in density forecasting. In: Mills, T. C., Patterson, K. (Eds.), Palgrave Handbook of
Econometrics, Volume 2: Applied Econometrics. Palgrave MacMillan, pp. 199–239.
Hamill, T. M., Colucci, S. J., 1997. Verification of Eta-RSM Short-Range Ensemble Forecasts. Monthly Weather Review 125 (6), 1312–
1327.
Hamilton, J. D., 1988. Rational-expectations econometric analysis of changes in regime: An investigation of the term structure of
interest rates. Journal of Economic Dynamics & Control 12 (2), 385–423.
Hamilton, J. D., 1990. Analysis of time series subject to changes in regime. Journal of Econometrics 45 (1), 39–70.
Hamilton, J. D., 2016. Macroeconomic regimes and regime shifts. In: Taylor, J. B., Uhlig, H. (Eds.), Handbook of Macroeconomics.
Vol. 2. Elsevier, pp. 163–201.
Hamilton, J. D., Lin, G., 1996. Stock market volatility and the business cycle. Journal of Applied Econometrics 11 (5), 573–593.
Han, J., Pei, J., Kamber, M., 2011. Data mining: concepts and techniques. Elsevier.
Han, P. K., Klein, W. M., Lehman, T. C., Massett, H., Lee, S. C., Freedman, A. N., 2009. Laypersons’ responses to the communication
of uncertainty regarding cancer risk estimates. Medical Decision Making 29 (3), 391–403.
Han, W., Wang, X., Petropoulos, F., Wang, J., 2019. Brain imaging and forecasting: Insights from judgmental model selection. Omega
87, 1–9.
Han, Y., He, A., Rapach, D. E., Zhou, G., 2020. Firm characteristics and expected stock returns. SSRN:3185335.
Hand, D. J., 2009. Mining the past to determine the future - problems and possibilities. International Journal of Forecasting 25 (3),
441–451.

202
Hanley, J. A., Joseph, L., Platt, R. W., Chung, M. K., Belisle, P., 2001. Visualizing the median as the minimum-deviation location. The
American Statistician 55 (2), 150–152.
Hannan, E. J., Quinn, B. G., 1979. The determination of the order of an autoregression. Journal of the Royal Statistical Society B, 41,
190–195.
Hansen, B. E., 2001. The New Econometrics of Structural Change: Dating breaks in US labour productivity. Journal of Economic
Perspectives 15 (4), 117–128.
Hansen, B. E., 2008. Least-squares forecast averaging. Journal of Econometrics 146 (2), 342–350.
Hansen, P. R., 2005. A test for superior predictive ability. Journal of Business & Economic Statistics 23 (4), 365–380.
Harford, T., 2014. Big data: A big mistake? Significance 11, 14–19.
Harrell, F. E., 2015. Regression modeling strategies: With applications to linear models, logistic and ordinal regression, and survival
analysis (2nd ed). New York, USA: Springer.
Harris, R. D., Yilmaz, F., 2010. Estimation of the conditional variance-covariance matrix of returns using the intraday range. Interna-
tional Journal of Forecasting 26 (1), 180–194.
Harvey, A. C., 1990. Forecasting, Structural Time Series Models and the Kalman Filter. Cambridge University Press.
Harvey, A. C., 2013. Dynamic Models for Volatility and Heavy Tails: With Applications to Financial and Economic Time Series.
Econometric Society Monographs. Cambridge University Press.
Harvey, C. R., Liu, Y., Zhu, H., 2016. ... and the cross-section of expected returns. Review of Financial Studies 29 (1), 5–68.
Harvey, D. I., Leybourne, S. J., Newbold, P., 1998. Tests for forecast encompassing. Journal of Business & Economic Statistics 16 (2),
254–259.
Harvey, N., 1995. Why are judgments less consistent in less predictable task situations? Organizational Behavior and Human Decision
Processes 63 (3), 247–263.
Harvey, N., 2007. Use of heuristics: Insights from forecasting research. Thinking & Reasoning 13 (1), 5–24.
Harvey, N., 2011. Anchoring and adjustment: A Bayesian heuristic? In: Brun, W., Keren, G., Kirkebøen, G., Montgomery, H. (Eds.),
Perspectives on Thinking, Judging, and Decision Making. Universitetsforlaget, Oslo, pp. 98–108.
Harvey, N., 2019. Commentary: Algorithmic aversion and judgmental wisdom. Foresight: The International Journal of Applied
Forecasting 54, 13–14.
Harvey, N., Bolger, F., 1996. Graphs versus tables: Effects of data presentation format on judgemental forecasting. International
Journal of Forecasting 12 (1), 119–137.
Harvey, N., Bolger, F., McClelland, A., 1994. On the nature of expectations. British Journal of Psychology 85 (2), 203–229.
Harvey, N., Reimers, S., 2013. Trend damping: Under-adjustment, experimental artifact, or adaptation to features of the natural
environment? Journal of Experimental Psychology. Learning, Memory, and Cognition 39 (2), 589–607.
Hasbrouck, J., 1995. One Security, Many Markets: Determining the Contributions to Price Discovery. Journal of Finance 50 (4),
1175–1199.
Hasni, M., Aguir, M. S., Babai, M. Z., Jemai, Z., 2019a. On the performance of adjusted bootstrapping methods for intermittent
demand forecasting. International Journal of Production Economics 216, 145–153.
Hasni, M., Aguir, M. S., Babai, M. Z., Jemai, Z., 2019b. Spare parts demand forecasting: a review on bootstrapping methods. Interna-
tional Journal of Production Research 57 (15-16), 4791–4804.
Hassan, S., Arroyo, J., Galán Ordax, J. M., Antunes, L., Pavón Mestras, J., 2013. Asking the oracle: Introducing forecasting principles
into agent-based modelling. Journal of Artificial Societies and Social Simulation 16 (3).
Hassani, H., Silva, E. S., 2015. Forecasting with big data: A review. Annals of Data Science 2, 5–19.
Hastie, T., Tibshirani, R., Friedman, J., 2009. The Elements of Statistical Learning. Springer-Verlag GmbH.
Hastie, T. J., Tibshirani, R. J., 1990. Generalized additive models. Vol. 43. CRC press.
Haugen, R. A., 2010. The New Finance, Overreaction, Complexity, and Their Consequences, 4th Edition. Pearson Education.
Haugen, R. A., Baker, N. L., 1996. Commonality in the determinants of expected stock returns. Journal of Financial Economics 41 (3),
401–439.
Hawkes, A. G., 1969. An approach to the analysis of electoral swing. Journal of the Royal Statistical Society, Series A 132 (1), 68–79.
Hawkes, A. G., 1971. Point spectra of some mutually exciting point processes. Journal of the Royal Statistical Society, Series B (Statis-
tical Methodology) 33 (3), 438–443.
Hawkes, A. G., 2018. Hawkes processes and their applications to finance: a review. Quantitative Finance 18 (2), 193–198.
Hawkes, A. G., Oakes, D., 1974. A cluster process representation of a Self-Exciting process. Journal of Applied Probability 11 (3),
493–503.
Hayes, B., 2002. Computing science: Statistics of deadly quarrels. American Scientist 90, 10–14.

203
He, A. W., Kwok, J. T., Wan, A. T., 2010. An empirical model of daily highs and lows of West Texas Intermediate crude oil prices.
Energy Economics 32 (6), 1499–1506.
He, K., Yu, L., Lai, K. K., 2012. Crude oil price analysis and forecasting using wavelet decomposed ensemble model. Energy 46 (1),
564–574.
Hecht, R., Gandhi, G., 2008. Demand forecasting for preventive AIDS vaccines. Pharmacoeconomics 26 (8), 679–697.
Hedonometer, 2020. Hedonometer word list. https://hedonometer.org/words/labMT-en-v2/, accessed on 2020-09-05.
Heinrich, C., 2014. The mode functional is not elicitable. Biometrika 101 (1), 245–251.
Heinrich, C., 2020. On the number of bins in a rank histogram. Quarterly Journal of the Royal Meteorological Society.
Heinrich, C., Hellton, K. H., Lenkoski, A., Thorarinsdottir, T. L., 2020. Multivariate postprocessing methods for high-dimensional
seasonal weather forecasts. Journal of the American Statistical Association.
Heligman, L., Pollard, J. H., 1980. The age pattern of mortality. Journal of the Institute of Actuaries 107, 49–80.
Hemri, S., 2018. Applications of postprocessing for hydrological forecasts. In: Statistical Postprocessing of Ensemble Forecasts. Else-
vier, pp. 219–240.
Hemri, S., Lisniak, D., Klein, B., 2015. Multivariate postprocessing techniques for probabilistic hydrological forecasting. Water Re-
sources Research 51 (9), 7436–7451.
Hendriks, F., Kienhues, D., Bromme, R., 2015. Measuring laypeople’s trust in experts in a digital age: The muenster epistemic
trustworthiness inventory (METI). PloS One 10 (10), e0139309.
Hendry, D., Clements, M., 2001. Forecasting Non-stationary Economic Time Series. MIT Press, Cambridge, Mass.
Hendry, D. F., 1995. Dynamic Econometrics. Oxford University Press, Oxford.
Hendry, D. F., 2001. Modelling UK inflation, 1875-1991. Journal of Applied Econometrics 16, 255–275.
Hendry, D. F., 2006. Robustifying Forecasts from Equilibrium-Correction Systems. Journal of Econometrics 135 (1-2), 399–426.
Hendry, D. F., 2010. Equilibrium-correction models. In: Macroeconometrics and Time Series Analysis. Springer, pp. 76–89.
Hendry, D. F. (Ed.), 2015. Introductory macro-econometrics: A new approach. Timberlake Consultants Press, London.
Hendry, D. F., 2020. First-in, first-out: Modelling the UK’s CO2 emissions, 1860–2016. Working paper 2020-W02, Nuffield College,
Oxford University.
Hendry, D. F., Doornik, J. A., 2014. Empirical Model Discovery and Theory Evaluation. MIT Press, Cambridge MA.
Hendry, D. F., Johansen, S., Santos, C., 2008. Automatic selection of indicators in a fully saturated regression. Computational Statistics
33, 317–335, erratum, 337–339.
Hendry, D. F., Mizon, G. E., 2012. Open-model forecast-error taxonomies. In: Chen, X., Swanson, N. R. (Eds.), Recent Advances and
Future Directions in Causality, Prediction, and Specification Analysis. Springer, pp. 219–240.
Herbst, E., Schorfheide, F., 2016. Bayesian Estimation of DSGE Models, 1st Edition. Princeton University Press.
Herrera, A. M., Hu, L., Pastor, D., 2018. Forecasting crude oil price volatility. International Journal of Forecasting 34 (4), 622–635.
Herrera, R., González, N., 2014. The modeling and forecasting of extreme events in electricity spot markets. International Journal of
Forecasting 30 (3), 477–490.
Herron, M. C., Shotts, K. W., 2004. Logical inconsistency in EI-Based Second-Stage regressions. American Journal of Political Science
48 (1), 172–183.
Hertzum, M., 2002. The importance of trust in software engineers’ assessment and choice of information sources. Information and
Organization 12 (1), 1–18.
Hertzum, M., 2014. Expertise seeking: A review. Information Processing & Management 50 (5), 775–795.
Hevia, C., Gonzalez-Rozada, M., Sola, M., Spagnolo, F., 2015. Estimating and Forecasting the Yield Curve Using A Markov Switching
Dynamic Nelson and Siegel Model. Journal of Applied Economics 30 (6), 987–1009.
Hewamalage, H., Bergmeir, C., Bandara, K., 2020. Recurrent neural networks for time series forecasting: Current status and future
directions. International Journal of Forecasting.
Hii, Y. L., Zhu, H., Ng, N., Ng, L. C., Rocklöv, J., 2012. Forecast of dengue incidence using temperature and rainfall. PLoS Neglected
Tropical Diseases 6 (11), e1908.
Hill, C. A., Zhang, G. P., Miller, K. E., 2018. Collaborative planning, forecasting, and replenishment & firm performance: An empirical
evaluation. International Journal of Production Economics 196, 12–23.
Hillebrand, E., Medeiros, M. C., 2010. The benefits of bagging for forecast models of realized volatility. Econometric Reviews 29 (5-6),
571–593.
Hinton, G. E., Srivastava, N., Krizhevsky, A., Sutskever, I., Salakhutdinov, R. R., 2012. Improving neural networks by preventing
co-adaptation of feature detectors. arXiv:1207.0580.
Hinton, Jr., H. L., 1999. Defence inventory, continuing challenger in managing inventories and avoiding adverse operational effects.

204
 Tech. rep., US General Accounting Office.
Hipel, K. W., McLeod, A. I., 1994. Time Series Modelling of Water Resources and Environmental Systems. Elsevier.
Hobijn, B., Franses, P. H., Ooms, M., 2004. Generalizations of the KPSS-test for stationarity. Statistica Neerlandica 58 (4), 483–502.
Hodges, P., Hogan, K., Peterson, J. R., Ang, A., 2017. Factor timing with cross-sectional and time-series predictors. Journal of Portfolio
Management 44 (1), 30–43.
Hodrick, R. J., Prescott, E. C., 1997. Postwar US business cycles: An empirical investigation. Journal of Money, Credit, and Banking,
1–16.
Hoerl, A. E., Kennard, R. W., 1970. Ridge regression: Biased estimation for nonorthogonal problems. Technometrics 12, 55–67.
Hoeting, J. A., Madigan, D., Raftery, A. E., Volinsky, C. T., 1999. Bayesian model averaging: A tutorial (with discussion). Statistical
Science 214, 382–417.
Hofmann, E., Rutschmann, E., 2018. Big data analytics and demand forecasting in supply chains: a conceptual analysis. The Interna-
tional Journal of Logistics Management 29 (2), 739–766.
Hogarth, R. M., Makridakis, S., 1981. Forecasting and planning: An evaluation. Management Science 27 (2), 115–138.
Holly, S., Pesaran, M. H., Yamagata, T., 2010. Spatial and temporal diffusion of house prices in the U.K. IZA Discussion Papers 4694,
Institute of Labor Economics (IZA).
Hong, T., Fan, S., 2016. Probabilistic electric load forecasting: A tutorial review. International Journal of Forecasting 32 (3), 914–938.
Hong, T., Pinson, P., 2019. Energy forecasting in the big data world. International Journal of Forecasting 35 (4), 1387–1388.
Hong, T., Pinson, P., Fan, S., 2014. Global energy forecasting competition 2012. International Journal of Forecasting 30 (2), 357–363.
Hong, T., Pinson, P., Fan, S., Zareipour, H., Troccoli, A., Hyndman, R. J., 2016. Probabilistic energy forecasting: Global energy
forecasting competition 2014 and beyond. International Journal of Forecasting 32 (3), 896–913.
Hong, T., Wang, P., Pahwa, A., Gui, M., Hsiang, S. M., 2010. Cost of temperature history data uncertainties in short term electric load
forecasting. In: 2010 IEEE 11th International Conference on Probabilistic Methods Applied to Power Systems. IEEE, pp. 212–217.
Hong, T., Xie, J., Black, J., 2019. Global energy forecasting competition 2017: Hierarchical probabilistic load forecasting. International
Journal of Forecasting 35 (4), 1389–1399.
Hong, W.-C., 2011. Traffic flow forecasting by seasonal SVR with chaotic simulated annealing algorithm. Neurocomputing 74 (12-13),
2096–2107.
Hong, Y., Li, H., Zhao, F., 2004. Out-of-Sample performance of Discrete-Time spot interest rate models. Journal of Business & Eco-
nomic Statistics 22 (4), 457–473.
Honnibal, M., 2015. spaCy: Industrial-strength Natural Language Processing (NLP) with Python and Cython. https://spacy.io,
accessed on 2020-09-10.
Honoré, C., Menut, L., Bessagnet, B., Meleux, F., Rouı̈l, L., Vautard, R., Poisson, N., Peuch, V., 2007. An integrated air quality forecast
system for a metropolitan area. Development in Environmental Science 6, 292–300.
Hooker, R. H., 1901. The suspension of the Berlin produce exchange and its effect upon corn prices. Journal of the Royal Statistical
Society 64 (4), 574–613.
Hora, S. C., 2004. Probability judgments for continuous quantities: Linear combinations and calibration. Management Science 50 (5),
597–604.
Hörmann, S., Horváth, L., Reeder, R., 2013. A functional version of the ARCH model. Econometric Theory 29 (2), 267–288.
Hornik, K., 1991. Approximation capabilities of multilayer feedforward networks. Neural Networks 4 (2), 251–257.
Hornik, K., Stinchcombe, M., White, H., 1989. Multilayer feedforward networks are universal approximators. Neural Networks 2 (5),
359–366.
Horrace, W. C., Schmidt, P., 2000. Multiple comparisons with the best, with economic applications. Journal of Applied Econometrics
15 (1), 1–26.
Horst, E. T., Rodriguez, A., Gzyl, H., Molina, G., 2012. Stochastic volatility models including open, close, high and low prices.
Quantitative Finance 12 (2), 199–212.
Horváth, L., Kokoszka, P., 2012. Inference for Functional Data with Applications. Springer, New York.
Horváth, L., Kokoszka, P., Rice, G., 2014. Testing stationarity of functional time series. Journal of Econometrics 179 (1), 66–82.
Horváth, L., Liu, Z., Rice, G., Wang, S., 2020. A functional time series analysis of forward curves derived from commodity futures.
International Journal of Forecasting 36 (2), 646–665.
Hoskins, B., 2013. The potential for skill across the range of the seamless weather-climate prediction problem: a stimulus for our
science. Quarterly Journal of the Royal Meteorological Society 139 (672), 573–584.
Hossin, M., Sulaiman, M., 2015. A review on evaluation metrics for data classification evaluations. International Journal of Data
Mining & Knowledge Management Process 5 (2), 1–11.

205
Hou, K., Xue, C., Zhang, L., 2020. Replicating anomalies. Review of Financial Studies 33 (5), 2019–2133.
Hou, Y., Edara, P., Sun, C., 2014. Traffic flow forecasting for urban work zones. IEEE transactions on intelligent transportation systems
16 (4), 1761–1770.
Hsu, J. C., 1981. Simultaneous confidence intervals for all distances from the “best”. The Annals of Statistics, 1026–1034.
Hu, K., Acimovic, J., Erize, F., Thomas, D. J., Van Mieghem, J. A., 2019. Forecasting new product life cycle curves: Practical approach
and empirical analysis. Manufacturing & Service Operations Management 21 (1), 66–85.
Huang, C., Chen, S., Yang, S., Kuo, C., 2015a. One-day-ahead hourly forecasting for photovoltaic power generation using an intelligent
method with weather-based forecasting models. IET Generation, Transmission and Distribution 9 (14), 1874–1882.
Huang, D., Jiang, F., Tu, J., Zhou, G., 2015b. Investor sentiment aligned: A powerful predictor of stock returns. Review of Financial
Studies 28 (3), 791–837.
Huang, J., Horowitz, J. L., Wei, F., 2010. Variable selection in nonparametric additive models. Annals of Statistics 38 (4), 2282–2313.
Huang, T., Fildes, R., Soopramanien, D., 2014. The value of competitive information in forecasting fmcg retail product sales and the
variable selection problem. European Journal of Operational Research 237 (2), 738–748.
Huang, T., Fildes, R., Soopramanien, D., 2019. Forecasting retailer product sales in the presence of structural change. European
Journal of Operational Research 279 (2), 459–470.
Huard, D., Évin, G., Favre, A.-C., 2006. Bayesian copula selection. Computational Statistics & Data Analysis 51 (2), 809–822.
Huber, J., Stuckenschmidt, H., 2020. Daily retail demand forecasting using machine learning with emphasis on calendric special
days. International Journal of Forecasting.
Huberty, M., 2015. Can we vote with our tweet? On the perennial difficulty of election forecasting with social media. International
Journal of Forecasting 31 (3), 992–1007.
Hubicka, K., Marcjasz, G., Weron, R., 2018. A note on averaging day-ahead electricity price forecasts across calibration windows.
IEEE Transactions on Sustainable Energy 10 (1), 321–323.
Hui, F. K. C., Warton, D. I., Foster, S. D., 2015. Tuning Parameter Selection for the Adaptive Lasso Using ERIC. Journal of the American
Statistical Society 110 (509), 262–269.
Hutwagner, L., Thompson, W., Seeman, G. M., Treadwell, T., 2003. The bioterrorism preparedness and response early aberration
reporting system (EARS). Journal of Urban Health 80 (1), i89–i96.
Hylleberg, S., 1994. Modelling seasonal variation. Nonstationary Time Series Analyses and Cointegration.
Hylleberg, S., Engle, R. F., Granger, C. W. J., Yoo, B. S., 1990. Seasonal integration and cointegration. Journal of Econometrics 44 (1),
215–238.
Hyndman, R., Athanasopoulos, G., Bergmeir, C., Caceres, G., Chhay, L., O’Hara-Wild, M., Petropoulos, F., Razbash, S., Wang, E.,
Yasmeen, F., 2020. forecast: Forecasting functions for time series and linear models. R package version 8.12.
Hyndman, R. J., 1996. Computing and graphing highest density regions. The American Statistician 50 (2), 120–126.
Hyndman, R. J., 2020. Quality measure for predictive highest density regions. Cross Validated, accessed on 2020-08-20.
URL https://stats.stackexchange.com/q/483882
Hyndman, R. J., Ahmed, R. A., Athanasopoulos, G., Shang, H. L., 2011. Optimal combination forecasts for hierarchical time series.
Computational Statistics and Data Analysis 55 (9), 2579–2589.
Hyndman, R. J., Athanasopoulos, G., 2018. Forecasting: principles and practice. OTexts.
Hyndman, R. J., Bashtannyk, D. M., Grunwald, G. K., 1996. Estimating and visualizing conditional densities. Journal of Computa-
tional and Graphical Statistics 5 (4), 315–336.
Hyndman, R. J., Billah, B., 2003. Unmasking the theta method. International Journal of Forecasting 19 (2), 287–290.
Hyndman, R. J., Koehler, A. B., Ord, J. K., Snyder, R. D., 2008. Forecasting with Exponential Smoothing: The State Space Approach.
Springer Verlag, Berlin.
Hyndman, R. J., Koehler, A. B., Snyder, R., Grose, S., 2002. A state space framework for automatic forecasting using exponential
smoothing methods. International Journal of Forecasting 18 (3), 439–454.
Hyndman, R. J., Shang, H. L., 2009. Forecasting functional time series (with discussions). Journal of the Korean Statistical Society
38 (3), 199–221.
Hyndman, R. J., Ullah, M. S., 2007. Robust forecasting of mortality and fertility rates: A functional data approach. Computational
Statistics & Data Analysis 51 (10), 4942–4956.
Ibrahim, R., L’Ecuyer, P., 2013. Forecasting call center arrivals: Fixed-effects, mixed-effects, and bivariate models. Manufacturing &
Service Operations Management 15 (1), 72–85.
Ibrahim, R., Ye, H., L’Ecuyer, P., Shen, H., 2016. Modeling and forecasting call center arrivals: A literature survey and a case study.
International Journal of Forecasting 32 (3), 865–874.

206
IEA, Paris, 2020. Electricity information: Overview.
URL www.iea.org/reports/electricity-information-overview
ifo Institute, 2020. ifo Business Climate Index for Germany. https://www.ifo.de/en/survey/ifo-business-climate-index, ac-
cessed on 2020-09-07.
IHME COVID-19 health service utilization forecasting team, Murray, C. J. L., 2020a. Forecasting COVID-19 impact on hospital bed-
days, ICU-days, ventilator-days and deaths by US state in the next 4 months. medrxiv;2020.03.27.20043752v1.
IHME COVID-19 health service utilization forecasting team, Murray, C. J. L., 2020b. Forecasting the impact of the first
wave of the COVID-19 pandemic on hospital demand and deaths for the USA and european economic area countries.
medrxiv;2020.04.21.20074732v1.
Ince, O., 2014. Forecasting exchange rates out-of-sample with panel methods and real-time data. Journal of International Money and
Finance 43 (C), 1–18.
Inoue, A., Jin, L., Rossi, B., 2017. Rolling window selection for out-of-sample forecasting with time-varying parameters. Journal of
Econometrics 196 (1), 55–67.
Inoue, A., Kilian, L., 2008. How useful is bagging in forecasting economic time series? A case study of us consumer price inflation.
Journal of the American Statistical Association 103 (482), 511–522.
Ioannidis, J. P. A., Cripps, S., Tanner, M. A., 2020. Forecasting for COVID-19 has failed. International Journal of Forecasting.
Irwin, G. A., Meeter, D. A., 1969. Building voter transition models from aggregate data. Midwest Journal of Political Science 13 (4),
545–566.
Islam, T., Meade, N., 2012. The impact of competition, and economic globalization on the multinational diffusion of 3G mobile
phones. Technological Forecasting and Social Change 79 (5), 843–850.
Jacobs, J. P. A. M., van Norden, S., 2011. Modeling data revisions: Measurement error and dynamics of ‘true’ values. Journal of
Econometrics 161, 101–109.
James, G., Witten, D., Hastie, T., Tibshirani, R., 2013. An Introduction to Statistical Learning with Applications in R. New York, USA:
Springer.
Jammazi, R., Aloui, C., 2012. Crude oil price forecasting: Experimental evidence from wavelet decomposition and neural network
modeling. Energy Economics 34 (3), 828–841.
Janczura, J., Trück, S., Weron, R., Wolff, R. C., 2013. Identifying spikes and seasonal components in electricity spot price data: A
guide to robust modeling. Energy Economics 38, 96–110.
Janke, T., Steinke, F., 2019. Forecasting the price distribution of continuous intraday electricity trading. Energies 12 (22), 4262.
Janssen, F., 2018. Advances in mortality forecasting: introduction. Genus 74 (21).
Januschowski, T., Arpin, D., Salinas, D., Flunkert, V., Gasthaus, J., Stella, L., Vazquez, P., 2018a. Now available in amazon Sage-
Maker: DeepAR algorithm for more accurate time series forecasting. https://aws.amazon.com/blogs/machine-learning/
now-available-in-amazon-sagemaker-deepar-algorithm-for-more-accurate-time-series-forecasting/, accessed on
2020-09-01.
Januschowski, T., Gasthaus, J., Wang, Y., Rangapuram, S. S., Callot, L., 2018b. Deep learning for forecasting: Current trends and
challenges. Foresight: The International Journal of Applied Forecasting 51, 42–47.
Januschowski, T., Gasthaus, J., Wang, Y., Salinas, D., Flunkert, V., Bohlke-Schneider, M., Callot, L., 2020. Criteria for classifying
forecasting methods. International Journal of Forecasting 36 (1), 167–177.
Januschowski, T., Kolassa, S., 2019. A classification of business forecasting problems. Foresight: The International Journal of Applied
Forecasting 52, 36–43.
Jardine, A. K. S., Lin, D., Banjevic, D., 2006. A review on machinery diagnostics and prognostics implementing condition-based
maintenance. Mechanical Systems and Signal Processing 20 (7), 1483–1510.
Jennings, W., Lewis-Beck, M., Wlezien, C., 2020. Election forecasting: Too far out? International Journal of Forecasting 36 (3), 949–
962.
Jeon, J., Panagiotelis, A., Petropoulos, F., 2019. Probabilistic forecast reconciliation with applications to wind power and electric load.
European Journal of Operational Research.
Jeon, J., Taylor, J., 2016. Short-term density forecasting of wave energy using ARMA-GARCH models and kernel density estimation.
International Journal of Forecasting 32 (3), 991–1004.
Jiang, J. J., Muhanna, W. A., Pick, R. A., 1996. The impact of model performance history information on users’ confidence in decision
models: An experimental examination. Computers in Human Behavior 12 (2), 193–207.
Jiao, E. X., Chen, J. L., 2019. Tourism forecasting: A review of methodological developments over the last decade. Tourism Economics
25 (3), 469–492.

207
Jing, G., Cai, W., Chen, H., Zhai, D., Cui, C., Yin, X., 2018. An air balancing method using support vector machine for a ventilation
system. Building and Environment 143, 487–495.
Joe, H., 1997. Multivariate models and dependence concepts. Chapman & Hall, London.
Joe, H., 2005. Asymptotic efficiency of the two-stage estimation method for copula-based models. Journal of Multivariate Analysis
94 (2), 401–419.
Joe, H., 2014. Dependence Modeling with Copulas. CRC Press.
Johansen, S., Nielsen, B., 2009a. An Analysis of the Indicator Saturation Estimator As a Robust Regression Estimator. In: Castle,
J., Shephard, N. (Eds.), The Methodology and Practice of Econometrics: A Festschrift in Honour of David F. Hendry. Oxford
University Press, Oxford and New York, pp. 1–35.
Johansen, S., Nielsen, B., 2009b. An analysis of the indicator saturation estimator as a robust regression estimator. In: Castle, J. L.,
Shephard, N. (Eds.), The Methodology and Practice of Econometrics: A Festschrift in Honour of David F. Hendry. Oxford University
Press, pp. 1–36.
Johnes, G., 1999. Forecasting unemployment. Applied Economics Letters 6 (9), 605–607.
Johnson, B. B., Slovic, P., 1995. Presenting uncertainty in health risk assessment: Initial studies of its effects on risk perception and
trust. Risk Analysis 15 (4), 485–494.
Johnston, D. M., 2008. The Historical Foundations of World Order: The Tower and the Arena. Martinus Nijhoff Publishers.
Johnston, R., Pattie, C., 2000. Ecological inference and Entropy-Maximizing: An alternative estimation procedure for Split-Ticket
voting. Political Analysis 8 (4), 333–345.
Johnstone, D. J., Jose, V. R. R., Winkler, R. L., 2011. Tailored scoring rules for probabilities. Decision Analysis 8 (4), 256–268.
Joiner, T. A., Leveson, L., Langfield-Smith, K., 2002. Technical language, advice understandability, and perceptions of expertise and
trustworthiness: The case of the financial planner. Australian Journal of Management 27 (1), 25–43.
Jondeau, E., 2007. Financial Modelling Under Non-Gaussian Distributions, 1st Edition. Springer, London.
Jongbloed, G., Koole, G., 2001. Managing uncertainty in call centres using poisson mixtures. Applied Stochastic Models in Business
and Industry 17 (4), 307–318.
Jonung, L., Larch, M., 2006. Improving fiscal policy in the EU: the case for independent forecasts. Economic Policy 21 (47), 491–534.
Jordá, O., Knüppelc, M., Marcellino, M., 2013. Empirical simultaneous prediction regions for path-forecasts. International Journal of
Forecasting 29 (3), 456–468.
Jore, A. S., Mitchell, J., Vahey, S. P., 2010. Combining forecast densities from VARs with uncertain instabilities. Journal of Applied
Econometrics 25 (4), 621–634.
Jose, V. R. R., Grushka-Cockayne, Y., Lichtendahl, K. C., 2014. Trimmed opinion pools and the crowd’s calibration problem. Manage-
ment Science 60 (2), 463–475.
Jose, V. R. R., Nau, R. F., Winkler, R. L., 2008. Scoring rules, generalized entropy, and utility maximization. Operations Research
56 (5), 1146–1157.
Jose, V. R. R., Winkler, R. L., 2008. Simple robust averages of forecasts: Some empirical results. International Journal of Forecasting
24 (1), 163–169.
Jose, V. R. R., Winkler, R. L., 2009. Evaluating quantile assessments. Operations Research 57 (5), 1287–1297.
Joslyn, S. L., Nadav-Greenberg, L., Taing, M. U., Nichols, R. M., 2009. The effects of wording on the understanding and use of
uncertainty information in a threshold forecasting decision. Applied Cognitive Psychology 23 (1), 55–72.
Joslyn, S. L., Nichols, R. M., 2009. Probability or frequency? expressing forecast uncertainty in public weather forecasts. Meteorolog-
ical Applications 16 (3), 309–314.
Julier, S. J., Uhlmann, J. K., 1997. New extension of the Kalman filter to nonlinear systems. In: Kadar, I. (Ed.), Signal Processing,
Sensor Fusion, and Target Recognition VI. Vol. 3068. International Society for Optics and Photonics, SPIE, pp. 182–193.
Kaastra, I., Boyd, M., 1996. Designing a neural network for forecasting financial and economic time series. Neurocomputing 10 (3),
215–236.
Kaboudan, M., 2001. Compumetric forecasting of crude oil prices. In: Proceedings of the 2001 Congress on Evolutionary Computa-
tion (IEEE Cat. No. 01TH8546). IEEE, pp. 283–287.
Kagraoka, Y., 2016. Common dynamic factors in driving commodity prices: Implications of a generalized dynamic factor model.
Economic Modelling 52, 609–617.
Kahn, K. B., 2002. An exploratory investigation of new product forecasting practices. Journal of Product Innovation Management
19 (2), 133–143.
Kahneman, D., 2011. Thinking, Fast and Slow. Penguin books, London.
Kahneman, D., Tversky, A., 1973. On the psychology of prediction. Psychological Review 80 (4), 237–251.

208
Kahneman, D., Tversky, A., 1996. On the reality of cognitive illusions. Psychological Review 103 (3), 582–91; discusion 592–6.
Kalamara, E., Turrell, A., Redl, C., Kapetanios, G., Kapadia, S., 2020. Making text count: economic forecasting using newspaper text.
Working Paper 865, Bank of England.
Kalman, R. E., 1960. A new approach to linear filtering and prediction problems. Journal of Fluids Engineering 82 (1), 35–45.
Kamarianakis, Y., Prastacos, P., 2005. Space–time modeling of traffic flow. Computers & Geosciences 31 (2), 119–133.
Kamisan, N. A. B., Lee, M. H., Suhartono, S., Hussin, A. G., Zubairi, Y. Z., 2018. Load forecasting using combination model of multiple
linear regression with neural network for Malaysian city. Sains Malaysiana 47 (2), 419–426.
Kang, S. H., Kang, S.-M., Yoon, S.-M., 2009. Forecasting volatility of crude oil markets. Energy Economics 31 (1), 119–125.
Kang, Y., 2012. Real-time change detection in time series based on growing feature quantization. In: The 2012 International Joint
Conference on Neural Networks (IJCNN). pp. 1–6.
Kang, Y., Belušić, D., Smith-Miles, K., 2014. Detecting and classifying events in noisy time series. Journal of the Atmospheric Sciences
71 (3), 1090–1104.
Kang, Y., Belušić, D., Smith-Miles, K., 2015. Classes of structures in the stable atmospheric boundary layer. Quarterly Journal of the
Royal Meteorological Society 141 (691), 2057–2069.
Kang, Y., Hyndman, R. J., Li, F., 2020. GRATIS: GeneRAting TIme Series with diverse and controllable characteristics. Statistical
Analysis and Data Mining 13 (4), 354–376.
Kang, Y., Hyndman, R. J., Smith-Miles, K., 2017. Visualising forecasting algorithm performance using time series instance spaces.
International Journal of Forecasting 33 (2), 345–358.
Kapetanios, G., Mitchell, J., Price, S., Fawcett, N., 2015. Generalised density forecast combinations. Journal of Econometrics 188 (1),
150–165.
Kargin, V., Onatski, A., 2008. Curve forecasting by functional autoregression. Journal of Multivariate Analysis 99 (10), 2508–2526.
Kascha, C., Ravazzolo, F., 2010. Combining inflation density forecasts. Journal of Forecasting 29 (1–2), 231–250.
Katz, R. W., Lazo, J. K. (Eds.), 2011. Economic Value of Weather and Climate Forecasts. The Oxford Handbook of Economic Forecast-
ing. Oxford University Press.
Kaufmann, R., Juselius, K., 2013. Testing hypotheses about glacial cycles against the observational record. Paleoceanography 28,
175–184.
Kawasaki, Y., Franses, P. H., 2004. Do seasonal unit roots matter for forecasting monthly industrial production? Journal of Forecasting
23 (2), 77–88.
Keane, M. P., Runkle, D. E., 1990. Testing the rationality of price forecasts: new evidence from panel data. American Economic Review
80 (4), 714–735.
Kedia, S., Williams, C., 2003. Predictors of substance abuse treatment outcomes in Tennessee. Journal of Drug Education 33 (1),
25–47.
Kelle, P., Silver, E. A., 1989. Forecasting the returns of reusable containers. Journal of Operations Management 8 (1), 17–35.
Kelly, B., Pruitt, S., 2013. Market expectations in the cross-section of present values. Journal of Finance 68 (5), 1721–1756.
Kennedy, R., Wojcik, S., Lazer, D., 2017. Improving election prediction internationally. Science 355 (6324), 515–520.
Kennedy, W. J., Wayne Patterson, J., Fredendall, L. D., 2002. An overview of recent literature on spare parts inventories. International
Journal of Production Economics 76 (2), 201–215.
Khaldi, R., El Afia, A., Chiheb, R., 2019. Forecasting of weekly patient visits to emergency department: real case study. Procedia
Computer Science 148, 532–541.
Kiesel, R., Paraschiv, F., 2017. Econometric analysis of 15-minute intraday electricity prices. Energy Economics 64, 77–90.
Kilian, L., Inoue, A., 2004. Bagging time series models. Tech. Rep. 110, Econometric Society.
Kilian, L., Taylor, M. P., 2003. Why is it so difficult to beat the random walk forecast of exchange rates? Journal of International
Economics 60 (1), 85–107.
Kim, C.-J., Kim Chang-Jin Nelson Charles, Nelson, C. R., 1999. State-Space Models with Regime Switching: Classical and Gibbs-
Sampling Approaches with Applications. MIT Press.
Kim, H. H., Swanson, N. R., 2014. Forecasting financial and macroeconomic variables using data reduction methods: New empirical
evidence. Journal of Econometrics 178, 352–367.
Kim, H. H., Swanson, N. R., 2018. Mining big data using parsimonious factor, machine learning, variable selection and shrinkage
methods. International Journal of Forecasting 34 (2), 339–354.
Kim, S., Shephard, N., Chib, S., 1998. Stochastic volatility: likelihood inference and comparison with ARCH models. Review of
Economic Studies 81, 361–393.
Kim, T. Y., Dekker, R., Heij, C., 2017. Spare part demand forecasting for consumer goods using installed base information. Computers

209
 & Industrial Engineering 103, 201–215.
King, G., 1997. A Solution to the Ecological Inference Problem: Reconstructing Individual Behavior from Aggregate Data. Princeton
University Press.
King, G., Rosen, O., Tanner, M. A., 1999. Binomial-Beta hierarchical models for ecological inference. Sociological Methods & Research
28 (1), 61–90.
King, G., Tanner, M. A., Rosen, O., 2004. Ecological Inference: New Methodological Strategies. Cambridge University Press.
Kingma, D. P., Ba, J., 2015. Adam: A method for stochastic optimization. Third Annual International Conference on Learning Repre-
sentations.
Kishor, N. K., Koenig, E. F., 2012. VAR estimation and forecasting when data are subject to revision. Journal of Business & Economic
Statistics 30 (2), 181–190.
Klepsch, J., Klüppelberg, C., 2017. An innovations algorithm for the prediction of functional linear processes. Journal of Multivariate
Analysis 155, 252–271.
Klepsch, J., Klüppelberg, C., Wei, T., 2017. Prediction of functional ARMA processes with an application to traffic data. Econometrics
and Statistics 1, 128–149.
Klima, A., Schlesinger, T., Thurner, P. W., Küchenhoff, H., 2019. Combining aggregate data and exit polls for the estimation of voter
transitions. Sociological Methods & Research 48 (2), 296–325.
Klima, A., Thurner, P. W., Molnar, C., Schlesinger, T., Küchenhoff, H., 2016. Estimation of voter transitions based on ecological
inference. AStA Advances in Statistical Analysis 2, 133–159.
Klofstad, C. A., Bishin, B. G., 2012. Exit and entrance polling: A comparison of election survey methods. Field Methods 24 (4),
429–437.
Koenig, E. F., Dolmas, S., Piger, J., 2003. The use and abuse of real-time data in economic forecasting. The Review of Economics and
Statistics 85 (3), 618–628.
Koenker, R., 2005. Quantile Regression. Econometric Society Monographs. Cambridge University Press.
Koh, Y.-M., Spindler, R., Sandgren, M., Jiang, J., 2018. A model comparison algorithm for increased forecast accuracy of dengue fever
incidence in Singapore and the auxiliary role of total precipitation information. International Journal of Environmental Health
Research 28 (5), 535–552.
Koirala, K. H., Mishra, A. K., D’Antoni, J. M., Mehlhorn, J. E., 2015. Energy prices and agricultural commodity prices: Testing
correlation using copulas method. Energy 81, 430–436.
Kokoszka, P., Reimherr, M., 2013. Determining the order of the functional autoregressive model. Journal of Time Series Analysis
34 (1), 116–129.
Kokoszka, P., Rice, G., Shang, H. L., 2017. Inference for the autocovariance of a functional time series under conditional heteroscedas-
ticity. Journal of Multivariate Analysis 162, 32–50.
Kolasa, M., Rubaszek, M., 2015a. Forecasting using DSGE models with financial frictions. International Journal of Forecasting 31 (1),
1–19.
Kolasa, M., Rubaszek, M., 2015b. How Frequently Should We Reestimate DSGE Models? International Journal of Central Banking
11 (4), 279–305.
Kolasa, M., Rubaszek, M., Skrzypczynśki, P., 2012. Putting the New Keynesian DSGE Model to the Real-Time Forecasting Test. Journal
of Money, Credit and Banking 44 (7), 1301–1324.
Kolassa, S., 2011. Combining exponential smoothing forecasts using Akaike weights. International Journal of Forecasting 27 (2),
238–251.
Kolassa, S., 2016. Evaluating predictive count data distributions in retail sales forecasting. International Journal of Forecasting 32 (3),
788–803.
Kolassa, S., 2020a. Quality measure for predictive Highest Density Regions. Cross Validated, accessed on 2020-08-20.
URL https://stats.stackexchange.com/q/483878
Kolassa, S., 2020b. Why the “best” point forecast depends on the error or accuracy measure. International Journal of Forecasting
36 (1), 208–211.
Kolassa, S., 2020c. Will deep and machine learning solve our forecasting problems? Foresight: The International Journal of Applied
Forecasting 57, 13–18.
Kolassa, S., Siemsen, E., 2016. Demand forecasting for managers. Business Expert Press.
Koning, A. J., Franses, P. H., Hibon, M., Stekler, H. O., 2005. The M3 competition: Statistical tests of the results. International Journal
of Forecasting 21 (3), 397–409.
Koop, G., Potter, S. M., 1999. Dynamic asymmetries in U.S. unemployment. Journal of Business & Economic Statistics 17 (3), 298–312.

210
Kostenko, A. V., Hyndman, R. J., 2006. A note on the categorization of demand patterns. Journal of the Operational Research Society
57 (10), 1256–1257.
Kotchoni, R., Leroux, M., Stevanovic, D., 2019. Macroeconomic forecast accuracy in a data-rich environment. Journal of Applied
Econometrics 34 (7), 1050–1072.
Kourentzes, N., Athanasopoulos, G., 2019. Cross-temporal coherent forecasts for Australian tourism. Annals of Tourism Research 75,
393–409.
Kourentzes, N., Athanasopoulos, G., 2020. Elucidate structure in intermittent demand series. European Journal of Operational Re-
search.
Kourentzes, N., Barrow, D., Petropoulos, F., 2019. Another look at forecast selection and combination: Evidence from forecast pooling.
International Journal of Production Economics 209, 226–235.
Kourentzes, N., Petropoulos, F., 2016. Forecasting with multivariate temporal aggregation: The case of promotional modelling. Inter-
national Journal of Production Economics 181, Part A, 145–153.
Kourentzes, N., Petropoulos, F., Trapero, J. R., 2014. Improving forecasting by estimating time series structural components across
multiple frequencies. International Journal of Forecasting 30 (2), 291–302.
Kourentzes, N., Rostami-Tabar, B., Barrow, D. K., 2017. Demand forecasting by temporal aggregation: Using optimal or multiple
aggregation levels? Journal of Business Research 78, 1–9.
Krishnan, T., Bass, F., Kummar, V., 2000. Impact of a late entrant on the diffusion of a new product/service. Journal of Marketing
Research 37, 269–278.
Krüger, E., Givoni, B., 2004. Predicting thermal performance in occupied dwellings. Energy and Buildings 36 (3), 301–307.
Krzysztofowicz, R., 1999. Bayesian theory of probabilistic forecasting via deterministic hydrologic model. Water Resources Research
35 (9), 2739–2750.
Krzysztofowicz, R., 2014. Probabilistic flood forecast: Exact and approximate predictive distributions. Journal of Hydrology 517,
643–651.
Kück, M., Crone, S. F., Freitag, M., 2016. Meta-learning with neural networks and landmarking for forecasting model selection an
empirical evaluation of different feature sets applied to industry data. In: 2016 International Joint Conference on Neural Networks
(IJCNN). IEEE, pp. 1499–1506.
Kuhn, M., Johnson, K., 2019. Feature Engineering and Selection. Taylor & Francis Ltd.
Kulkarni, G., Kannan, P. K., Moe, W., 2012. Using online search data to forecast new product sales. Decision Support Systems 52 (3),
604–611.
Kumar, D., 2015. Sudden changes in extreme value volatility estimator: Modeling and forecasting with economic significance analysis.
Economic Modelling 49, 354–371.
Künsch, H. R., 1989. The jackknife and the bootstrap for general stationary observations. Annals of Statistics 17 (3), 1217–1241.
Kuster, C., Rezgui, Y., Mourshed, M., 2017. Electrical load forecasting models: A critical systematic review. Sustainable Cities and
Society 35, 257–270.
Kusters, U., McCullough, B., Bell, M., 2006. Forecasting software: Past, present and future. International Journal of Forecasting 22 (3),
599–615.
Kwiatkowski, D., Phillips, P. C. B., Schmidt, P., Shin, Y., 1992. Testing the null hypothesis of stationarity against the alternative of a
unit root: How sure are we that economic time series have a unit root? Journal of Econometrics 54 (1), 159–178.
Kyriazi, F., Thomakos, D. D., Guerard, J. B., 2019. Adaptive learning forecasting, with applications in forecasting agricultural prices.
International Journal of Forecasting 35 (4), 1356–1369.
La Fabrique des Mobilités, 2020. Motorway traffic in Luxembourg. Accessed on 2020-09-01.
URL https://www.kaggle.com/fabmob/motorway-traffic-in-luxembourg?select=datexDataA1.csv
Labarere, J., Bertrand, R., Fine, M. J., 2014. How to derive and validate clinical prediction models for use in intensive care medicine.
Intensive Care Medicine 40 (4), 513–527.
Ladiray, D., Quenneville, B., 2001. Seasonal Adjustment with the X-11 Method. Lecture Notes in Statistics 158. Springer, New York,
USA.
Lago, J., De Ridder, F., De Schutter, B., 2018. Forecasting spot electricity prices: Deep learning approaches and empirical comparison
of traditional algorithms. Applied Energy 221, 386–405.
Lahiri, S. K., Lahiri, N., 2003. Resampling Methods for Dependent Data (Springer Series in Statistics). Springer.
Lai, G., Chang, W.-C., Yang, Y., Liu, H., 2018. Modeling long-and short-term temporal patterns with deep neural networks. In: The
41st International ACM SIGIR Conference on Research & Development in Information Retrieval. pp. 95–104.
Landon, J., Ruggeri, F., Soyer, R., Tarimcilar, M. M., 2010. Modeling latent sources in call center arrival data. European Journal of

211
 Operational Research 204 (3), 597–603.
Lanne, M., Saikkonen, P., 2003. Modeling the U.S. Short-Term interest rate by mixture autoregressive processes. Journal of Financial
Econometrics 1 (1), 96–125.
Larrick, R. P., Soll, J. B., 2006. Intuitions about combining opinions: Misappreciation of the averaging principle. Management Science
52 (1), 111–127.
Larson, P. D., Simchi-Levi, D., Kaminsky, P., Simchi-Levi, E., 2001. Designing and managing the supply chain: Concepts, strategies,
and case studies. Journal of Business Logistics 22 (1), 259–261.
Law, R., Li, G., Fong, D. K. C., Han, X., 2019. Tourism demand forecasting: A deep learning approach. Annals of Tourism Research
75, 410–423.
Lawrence, M., 2000. What does it take to achieve adoption in sales forecasting? International Journal of Forecasting 16 (2), 147–148.
Lawrence, M., Goodwin, P., Fildes, R., 2002. Influence of user participation on DSS use and decision accuracy. Omega 30 (5), 381–392.
Lawrence, M., Goodwin, P., O’Connor, M., Önkal, D., 2006. Judgmental forecasting: A review of progress over the last 25 years.
International Journal of Forecasting 22 (3), 493–518.
Lawrence, M., Makridakis, S., 1989. Factors affecting judgmental forecasts and confidence intervals. Organizational Behavior and
Human Decision Processes 43 (2), 172–187.
Lawrence, M., O’Connor, M., 1992. Exploring judgemental forecasting. International Journal of Forecasting 8 (1), 15–26.
Layard, R., Nickell, S. J., Jackman, R., 1991. Unemployment, Macroeconomic Performance and the Labour Market. Oxford University
Press, Oxford.
Le, Q., Mikolov, T., 2014. Distributed representations of sentences and documents. In: International Conference on Machine Learning.
pp. 1188–1196.
Leadbetter, M. R., 1991. On a basis for ‘peaks over threshold’ modeling. Statistics & Probability Letters 12 (4), 357–362.
Leal, T., Pérez, J. J., Tujula, M., Vidal, J. P., 2008. Fiscal forecasting: Lessons from the literature and challenges. Fiscal Studies 29,
347–386.
Ledolter, J., 1989. The effect of additive outliers on the forecasts from ARIMA models. International Journal of Forecasting 5 (2),
231–240.
Ledolter, J., 1991. Outliers in time series analysis: Some comments on their impact and their detection. Image.
Lee, H. L., Padmanabhan, V., Whang, S., 2004. Information distortion in a supply chain: The bullwhip effect. Management Science
50, 1875–1886.
Lee, J., Milesi-Ferretti, G. M., Ricci, L. A., 2013. Real exchange rates and fundamentals: A cross-country perspective. Journal of
Money, Credit and Banking 45 (5), 845–865.
Lee, R. D., Carter, L. R., 1992. Modeling and forecasting US mortality. Journal of the American Statistical Association 87 (419),
659–671.
Lee, W. Y., Goodwin, P., Fildes, R., Nikolopoulos, K., Lawrence, M., 2007. Providing support for the use of analogies in demand
forecasting tasks. International Journal of Forecasting 23 (3), 377–390.
Leigh, C., Alsibai, O., Hyndman, R. J., Kandanaarachchi, S., King, O. C., McGree, J. M., Neelamraju, C., Strauss, J., Talagala, P. D.,
Turner, R. D., Kerrie, M., Peterson, E. E., 2019. A framework for automated anomaly detection in high frequency water-quality
data from in situ sensors. Science of The Total Environment 664, 885–898.
Lemke, C., Gabrys, B., 2010. Meta-learning for time series forecasting and forecast combination. Neurocomputing 73 (10-12), 2006–
2016.
Lemke, W., Archontakis, T., 2008. Bond pricing when the short-term interest rate follows a threshold process. Quantitative Finance
8 (8), 811–822.
Lerch, S., Baran, S., Möller, A., Groß, J., Schefzik, R., Hemri, S., Graeter, M., 2020. Simulation-based comparison of multivariate
ensemble post-processing methods. Nonlinear Processes in Geophysics 27 (2), 349–371.
Leuenberger, D., Haefele, A., Omanovic, N., Fengler, M., Martucci, G., Calpini, B., Fuhrer, O., Rossa, A., 2020. Improving high-
impact numerical weather prediction with lidar and drone observations. Bulletin of the American Meteorological Society 101 (7),
E1036–E1051.
Leva, S., Mussetta, M., Ogliari, E., 2019. PV module fault diagnosis based on microconverters and Day-Ahead forecast. IEEE Transac-
tions on Industrial Electronics 66 (5), 3928–3937.
Levine, R., Pickett, J., Sekhri, N., Yadav, P., 2008. Demand forecasting for essential medical technologies. American Journal of Law &
Medicine 34 (2-3), 225–255.
Lewellen, J., 2015. The cross-section of expected stock returns. Critical Finance Review 4 (1), 1–44.
Lewis, B., Herbert, R., Bell, R., 2003. The application of fourier analysis to forecasting the inbound call time series of a call centre.

212
 In: Proceedings of the International Congress on Modeling and Simulation (MODSIM03); Townsville, Australia. Citeseer, pp.
1281–1286.
Lewis-Beck, M. S., 2005. Election forecasting: Principles and practice. British Journal of Politics and International Relations 7 (2),
145–164.
Li, D., Robinson, P. M., Shang, H. L., 2020a. Local Whittle estimation of long range dependence for functional time series. Working
paper, University of York.
Li, D., Robinson, P. M., Shang, H. L., 2020b. Long-range dependent curve time series. Journal of the American Statistical Association
115 (530), 957–971.
Li, D., Robinson, P. M., Shang, H. L., 2020c. Nonstationary fractionally integrated functional time series. Working paper, University
of York.
Li, D. X., 2000. On default correlation: A copula function approach. The Journal of Fixed Income 9 (4), 43–54.
Li, F., He, Z., 2019. Credit risk clustering in a business group: which matters more, systematic or idiosyncratic risk? Cogent Economics
& Finance, 1632528.
Li, F., Kang, Y., 2018. Improving forecasting performance using covariate-dependent copula models. International Journal of Fore-
casting 34 (3), 456–476.
Li, G., Jiao, E., 2020. Tourism forecasting research: a perspective article. Tourism Review.
Li, G., Song, H., Witt, S. F., 2005. Recent developments in econometric modeling and forecasting. Journal of Travel Research 44 (1),
82–99.
Li, H., Hong, Y., 2011. Financial volatility forecasting with range-based autoregressive volatility model. Finance Research Letters
8 (2), 69–76.
Li, J., Li, G., Liu, M., Zhu, X., Wei, L., 2020d. A novel text-based framework for forecasting agricultural futures using massive online
news headlines. International Journal of Forecasting.
Li, J., Liao, Z., Quaedvlieg, R., 2020e. Conditional superior predictive ability. SSRN:3536461.
Li, J. S.-H., Chan, W.-S., 2011. Time-simultaneous prediction bands: A new look at the uncertainty involved in forecasting mortality.
Insurance: Mathematics and Economics 49 (1), 81–88.
Li, L., Noorian, F., Moss, D. J., Leong, P. H., 2014. Rolling window time series prediction using MapReduce. In: Proceedings of the
2014 IEEE 15th International Conference on Information Reuse and Integration (IEEE IRI 2014). IEEE, pp. 757–764.
Li, M., Huang, L., Gong, L., 2011. Research on the challenges and solutions of design large-scale call center intelligent scheduling
system. Procedia Engineering 15, 2359–2363.
Li, W., Han, Z., Li, F., 2008. Clustering analysis of power load forecasting based on improved ant colony algorithm. In: 2008 7th
World Congress on Intelligent Control and Automation. pp. 7492–7495.
Li, X., Kang, Y., Li, F., 2020f. Forecasting with time series imaging. Expert System with Applications 160, 113680.
Liberty, E., Karnin, Z., Xiang, B., Rouesnel, L., Coskun, B., Nallapati, R., Delgado, J., Sadoughi, A., Astashonok, A., Das, P., Balioglu,
C., Charkravarty, S., Jha, M., Gaultier, P., Januschowski, T., Flunkert, V., Wang, B., Gasthaus, J., Rangapuram, S., Salinas, D.,
Schelter, S., Arpin, D., Smola, A., 2020. Elastic machine learning algorithms in Amazon SageMaker. In: Proceedings of the 2020
International Conference on Management of Data. SIGMOD ’20. ACM, New York, NY, USA, pp. 731–737.
Lichtendahl, K. C., Grushka-Cockayne, Y., Winkler, R. L., 2013. Is it better to average probabilities or quantiles? Management Science
59 (7), 1594–1611.
Lichtendahl Jr, K. C., Winkler, R. L., 2020. Why do some combinations perform better than others? International Journal of Forecast-
ing 36 (1), 142–149.
Lildholdt, P. M., 2002. Estimation of GARCH models based on open, close, high, and low prices, Centre for Analytical Finance, Aarhus
School of Business.
Lim, J. S., O’Connor, M., 1996a. Judgmental forecasting with interactive forecasting support systems. Decision Support Systems
16 (4), 339–357.
Lim, J. S., O’Connor, M., 1996b. Judgmental forecasting with time series and causal information. International Journal of Forecasting
12 (1), 139–153.
Limaye, V. S., Vargo, J., Harkey, M., Holloway, T., Patz, J. A., 2018. Climate change and heat-related excess mortality in the eastern
usa. EcoHealth 15 (3), 485–496.
Lin, C.-F. J., Teräsvirta, T., 1994. Testing the constancy of regression parameters against continuous structural change. Journal of
Econometrics 62 (2), 211–228.
Lin, E. M., Chen, C. W., Gerlach, R., 2012. Forecasting volatility with asymmetric smooth transition dynamic range models. Interna-
tional Journal of Forecasting 28 (2), 384–399.

213
Ling, S., 1999. On the probabilistic properties of a double threshold ARMA conditional heteroskedastic model. Journal of Applied
Probability 36 (3), 688–705.
Ling, S., Tong, H., Li, D., 2007. Ergodicity and invertibility of threshold Moving-Average models. Bernoulli 13 (1), 161–168.
Linnér, L., Eriksson, I., Persson, M., Wettermark, B., 2020. Forecasting drug utilization and expenditure: ten years of experience in
stockholm. BMC Health Services Research 20, 1–11.
Litsiou, K., Polychronakis, Y., Karami, A., Nikolopoulos, K., 2019. Relative performance of judgmental methods for forecasting the
success of megaprojects. International Journal of Forecasting.
Liu, W., Zhu, F., Zhao, T., Wang, H., Lei, X., Zhong, P.-A., Fthenakis, V., 2020. Optimal stochastic scheduling of hydropower-based
compensation for combined wind and photovoltaic power outputs. Applied Energy 276, 115501.
Liu, Y., 2005. Value-at-Risk model combination using artificial neural networks. Ermory University Working Papers.
Ljung, G. M., Box, G. E., 1978. On a measure of lack of fit in time series models. Biometrika 65 (2), 297–303.
Loaiza-Maya, R., Smith, M. S., 2020. Real-time macroeconomic forecasting with a heteroscedastic inversion copula. Journal of Busi-
ness & Economic Statistics 38 (2), 470–486.
Locarek-Junge, H., Prinzler, R., 1998. Estimating Value-at-Risk using neural networks. In: Informationssysteme in der Fi-
nanzwirtschaft. Springer Berlin Heidelberg, pp. 385–397.
Logg, J. M., Minson, J. A., Moore, D. A., 2019. Algorithm appreciation: People prefer algorithmic to human judgment. Organizational
Behavior and Human Decision Processes 151, 90–103.
Lohmann, T., Hering, A. S., Rebennack, S., 2016. Spatio-temporal hydro forecasting of multireservoir inflows for hydro-thermal
scheduling. European Journal of Operational Research 255 (1), 243–258.
Lombardo, J. S., Burkom, H., Pavlin, J., 2004. ESSENCE II and the framework for evaluating syndromic surveillance systems. Mor-
bidity and Mortality Weekly Report, 159–165.
Loper, E., Bird, S., 2002. NLTK: the natural language toolkit. arXiv:cs/0205028.
López, C., Zhong, W., Zheng, M., 2017. Short-term electric load forecasting based on wavelet neural network, particle swarm opti-
mization and ensemble empirical mode decomposition. Energy Procedia 105, 3677–3682.
López, M., Valero, S., Senabre, C., Aparicio, J., Gabaldon, A., 2012. Application of SOM neural networks to short-term load forecast-
ing: The spanish electricity market case study. Electric Power Systems Research 91, 18–27.
López Cabrera, B., Schulz, F., 2016. Volatility linkages between energy and agricultural commodity prices. Energy Economics 54 (C),
190–203.
López-Ruiz, A., Bergillos, R. J., Ortega-Sánchez, M., 2016. The importance of wave climate forecasting on the decision-making process
for nearshore wave energy exploitation. Applied Energy 182, 191–203.
Lopez-Suarez, C. F., Rodriguez-Lopez, J. A., 2011. Nonlinear exchange rate predictability. Journal of International Money and Finance
30 (5), 877–895.
Lothian, J. R., Taylor, M. P., 1996. Real exchange rate behavior: The recent float from the perspective of the past two centuries. Journal
of Political Economy 104 (3), 488–509.
Lotka, A., 1920. Undamped oscillations derived from the law of mass action. Journal of the American Chemical Society 42, 1595–99.
Lotka, A. J., 1925. Elements of Physical Biology. Williams & Wilkins.
Lotze, T. H., Shmueli, G., 2009. How does improved forecasting benefit detection? An application to biosurveillance. International
Journal of Forecasting 25 (3), 467–483.
Lovins, JB, 1968. Development of a stemming algorithm. Mechanical Translation and Computational Linguistics 11 (1-2), 22–31.
Lowe, R., Bailey, T. C., Stephenson, D. B., Graham, R. J., Coelho, C. A., Carvalho, M. S., Barcellos, C., 2011. Spatio-temporal modelling
of climate-sensitive disease risk: Towards an early warning system for dengue in Brazil. Computers & Geosciences 37 (3), 371–381.
Lu, H., Azimi, M., Iseley, T., 2019. Short-term load forecasting of urban gas using a hybrid model based on improved fruit fly opti-
mization algorithm and support vector machine. Energy Reports 5, 666–677.
Lu, X., Dong, Z. Y., Li, X., 2005. Electricity market price spike forecast with data mining techniques. Electric Power Systems Research
73 (1), 19–29.
Lübbers, J., Posch, P. N., 2016. Commodities’ common factor: An empirical assessment of the markets’ drivers. Journal of Commodity
Markets 4 (1), 28–40.
Lucas, A., Schwaab, B., Zhang, X., 2014. Conditional euro area sovereign default risk. Journal of Business & Economic Statistics 32 (2),
271–284.
Lucas, R. E., 1976. Econometric policy evaluation: A critique. Carnegie-Rochester Conference Series on Public Policy 1, 19–46.
Ludvigson, S. C., Ng, S., 2007. The empirical risk-return relation: A factor analysis approach. Journal of Financial Economics 83 (1),
171–222.

214
Luo, J., Hong, T., Fang, S.-C., 2018a. Benchmarking robustness of load forecasting models under data integrity attacks. International
Journal of Forecasting 34 (1), 89–104.
Luo, J., Hong, T., Yue, M., 2018b. Real-time anomaly detection for very short-term load forecasting. Journal of Modern Power Systems
and Clean Energy 6 (2), 235–243.
Luo, J., Klein, T., Ji, Q., Hou, C., 2019. Forecasting realized volatility of agricultural commodity futures with infinite Hidden Markov
HAR models. International Journal of Forecasting.
Lütkepohl, H., 2005. New Introduction to Multiple Time Series Analysis. Springer, Berlin, Heidelberg.
Lütkepohl, H., 2011. Forecasting nonlinear aggregates and aggregates with time-varying weights. Jahrbücher für Nationalökonomie
und Statistik 231 (1), 107–133.
Lux, T., 2008. The Markov-Switching multifractal model of asset returns. Journal of Business & Economic Statistics 26 (2), 194–210.
L’heureux, A., Grolinger, K., Elyamany, H. F., Capretz, M. A., 2017. Machine learning with big data: Challenges and approaches. IEEE
Access 5, 7776–7797.
Ma, F., Chitta, R., Zhou, J., You, Q., Sun, T., Gao, J., 2017. Dipole: Diagnosis prediction in healthcare via attention-based bidirectional
recurrent neural networks. In: Proceedings of the 23rd ACM SIGKDD international conference on knowledge discovery and data
mining. pp. 1903–1911.
Ma, S., 2021. A hybrid deep meta-ensemble networks with application in electric utility industry load forecasting. Information
Sciences 544, 183–196.
Ma, S., Fildes, R., 2017. A retail store SKU promotions optimization model for category multi-period profit maximization. European
Journal of Operational Research 260 (2), 680–692.
Ma, S., Fildes, R., Huang, T., 2016. Demand forecasting with high dimensional data: The case of SKU retail sales forecasting with
intra-and inter-category promotional information. European Journal of Operational Research 249 (1), 245–257.
Macaulay, F. R., 1931. The smoothing of time series. NBER Books.
MacDonald, R., 1998. What determines real exchange rates? The long and the short of it. Journal of International Financial Markets,
Institutions and Money 8 (2), 117–153.
MacDonald, R., Marsh, I. W., 1994. Combining exchange rate forecasts: What is the optimal consensus measure? Journal of Forecast-
ing 13 (3), 313–332.
Madaus, L., McDermott, P., Hacker, J., Pullen, J., 2020. Hyper-local, efficient extreme heat projection and analysis using machine
learning to augment a hybrid dynamical-statistical downscaling technique. Urban Climate 32, 100606.
Maddix, D. C., Wang, Y., Smola, A., 2018. Deep Factors with Gaussian Processes for Forecasting. arXiv:1812.00098.
Madhavan, P., Wiegmann, D. A., 2007. Similarities and differences between human–human and human–automation trust: an inte-
grative review. Theoretical Issues in Ergonomics Science 8 (4), 277–301.
Mahajan, V., Muller, E., Bass, F., 1990. New product diffusion models in marketing: a review and directions of future research. Journal
of Marketing 54, 1–26.
Maheu, J. M., Yang, Q., 2016. An infinite hidden Markov model for short-term interest rates. Journal of Empirical Finance 38, 202–
220.
Maister, D. H., Galford, R., Green, C., 2012. The Trusted Advisor. Simon and Schuster.
Makridakis, S., Andersen, A., Carbone, R., Fildes, R., Hibon, M., Lewandowski, R., Newton, J., Parzen, E., Winkler, R., 1982. The
accuracy of extrapolation (time series) methods: Results of a forecasting competition. Journal of Forecasting 1 (2), 111–153.
Makridakis, S., Bonneli, E., Clarke, S., Fildes, R., Gilliland, M., Hover, J., Tashman, J., 2020a. The benefits of systematic forecasting
for organizations: The UFO project. Foresight: The International Journal of Applied Forecasting 59, 45–56.
Makridakis, S., Chatfield, C., Hibon, M., Lawrence, M., Mills, T., Ord, K., Simmons, L. F., 1993. The M2-competition: A real-time
judgmentally based forecasting study. International Journal of Forecasting 9 (1), 5–22.
Makridakis, S., Hibon, M., 1979. Accuracy of forecasting: An empirical investigation. Journal of the Royal Statistical Society: Series
A (General) 142 (2), 97–125.
Makridakis, S., Hibon, M., 2000. The M3-Competition: results, conclusions and implications. International Journal of Forecasting
16 (4), 451–476.
Makridakis, S., Hyndman, R. J., Petropoulos, F., 2020b. Forecasting in social settings: The state of the art. International Journal of
Forecasting 36 (1), 15–28.
Makridakis, S., Kirkham, R., Wakefield, A., Papadaki, M., Kirkham, J., Long, L., 2019. Forecasting, uncertainty and risk; perspectives
on clinical decision-making in preventive and curative medicine. International Journal of Forecasting 35 (2), 659–666.
Makridakis, S., Spiliotis, E., Assimakopoulos, V., 2018. Statistical and machine learning forecasting methods: Concerns and ways
forward. PLoS One 13 (3), 1–26.

215
Makridakis, S., Spiliotis, E., Assimakopoulos, V., 2020c. The M4 competition: 100,000 time series and 61 forecasting methods. Inter-
national Journal of Forecasting 36 (1), 54–74.
Makridakis, S., Spiliotis, E., Assimakopoulos, V., 2020d. The M5 Accuracy competition: Results, findings and conclusions. Working
paper, National Technical University of Athens.
Makridakis, S., Winkler, R. L., 1989. Sampling distributions of post-sample forecasting errors. Journal of the Royal Statistical Society:
Series C 38 (2), 331–342.
Makridakis, S. G., Hogarth, R. M., Gaba, A., 2010. Dance With Chance: Making Luck Work for You. Oneworld Publications.
Mamdani, E. H., Assilian, S., 1975. An experiment in linguistic synthesis with a fuzzy logic controller. International Journal of Man-
Machine Studies 7, 1–15.
Mandal, P., Madhira, S. T. S., Haque, A. U., Meng, J., Pineda, R. L., 2012. Forecasting power output of solar photovoltaic system using
wavelet transform and artificial intelligence techniques. Procedia Computer Science 12, 332–337.
Mandelbrot, B., 1963. The Variation of Certain Speculative Prices. The Journal of Business 36 (4), 394.
Mandelbrot, B. B., 1983. The Fractal Geometry of Nature. Henry Holt and Company.
Manders, A., Schaap, M., Hoogerbrugge, R., 2009. Testing the capability of the chemistry transport model LOTOS-EUROS to forecast
PM10 levels in the Netherlands. Atmospheric Environment 46, 4050–4059.
Mankiw, N. G., Reis, R., 2002. Sticky information versus sticky prices: A proposal to replace the New Keynesian Phillips Curve.
Quarterly Journal of Economics 117, 1295–1328.
Mankiw, N. G., Reis, R., Wolfers, J., 2003. Disagreement about inflation expectations. Tech. rep., National Bureau of Economic Re-
search, Cambridge MA.
Mann, M., 2018. Have wars and violence declined? Theory and Society 47 (1), 37–60.
Manna, S., Biswas, S., Kundu, R., Rakshit, S., Gupta, P., Barman, S., 2017. A statistical approach to predict flight delay using gradient
boosted decision tree. In: 2017 International Conference on Computational Intelligence in Data Science (ICCIDS). IEEE, pp. 1–5.
Manner, H., Türk, D., Eichler, M., 2016. Modeling and forecasting multivariate electricity price spikes. Energy Economics 60, 255–
265.
Mannes, A. E., Larrick, R. P., Soll, J. B., 2012. The social psychology of the wisdom of crowds. Social Judgment and Decision Making.
297, 227–242.
Mannes, A. E., Soll, J. B., Larrick, R. P., 2014. The wisdom of select crowds. Journal of Personality and Social Psychology 107 (2),
276–299.
Manning, C., Schütze, H., Raghavan, P., 2008. Introduction to information retrieval. Cambridge University Press.
Manski, C. F., Molinari, F., 2010. Rounding probabilistic expectations in surveys. Journal of Business & Economic Statistics 28:2,
219–231.
Mapa, D., 2003. A range-based GARCH model for forecasting volatility. The Philippine Review of Economics 60 (2), 73–90.
Marangon Lima, L. M., Popova, E., Damien, P., 2014. Modeling and forecasting of Brazilian reservoir inflows via dynamic linear
models. International Journal of Forecasting 30 (3), 464–476.
Marchetti, C., 1983. The automobile in a system context: The past 80 years and the next 20 years. Technological Forecasting and
Social Change 23 (1), 3–23.
Marchetti, C., Nakicenovic, N., 1979. The dynamics of energy systems and the logistic substitution model. International Institute for
Applied Systems Analysis, Laxenburg, Austria RR-79-13, 1–71.
Marcjasz, G., Uniejewski, B., Weron, R., 2019. On the importance of the long-term seasonal component in day-ahead electricity price
forecasting with NARX neural networks. International Journal of Forecasting 35 (4), 1520–1532.
Marcjasz, G., Uniejewski, B., Weron, R., 2020. Beating the naı̈ve—combining LASSO with naı̈ve intraday electricity price forecasts.
Energies 13 (7), 1667.
Marczak, M., Proietti, T., 2016. Outlier detection in structural time series models: The indicator saturation approach. International
Journal of Forecasting 32 (1), 180–202.
Mark, N. C., 1995. Exchange rates and fundamentals: Evidence on long-horizon predictability. American Economic Review 85 (1),
201–218.
Mark, N. C., Sul, D., 2001. Nominal exchange rates and monetary fundamentals: Evidence from a small post-Bretton Woods panel.
Journal of International Economics 53 (1), 29–52.
Markowitz, H., 1952. Portfolio Selection. The Journal of Finance 7 (1), 77–91.
Marron, J. S., Wand, M. P., 1992. Exact mean integrated squared error. Annals of Statistics 20 (2), 712–736.
Martinez, A. B., Castle, J. L., Hendry, D. F., 2019. Smooth robust multi-step forecasting methods. Working paper, Nuffield College,
Oxford University.

216
Martinez, E. Z., Silva, E. A. S. d., 2011. Predicting the number of cases of dengue infection in Ribeirão Preto, São Paulo State, Brazil,
using a SARIMA model. Cadernos de Saude Publica 27, 1809–1818.
Martı́nez-Álvarez, F., Troncoso, A., Riquelme, J. C., Aguilar-Ruiz, J. S., 2011. Discovery of motifs to forecast outlier occurrence in time
series. Pattern Recognition Letters 32 (12), 1652–1665.
Martinez Alvarez, F., Troncoso, A., Riquelme, J. C., Aguilar Ruiz, J. S., 2011. Energy time series forecasting based on pattern sequence
similarity. IEEE Transactions on Knowledge and Data Engineering 23 (8), 1230–1243.
Masarotto, G., 1990. Bootstrap prediction intervals for autoregressions. International Journal of Forecasting 6 (2), 229–239.
Mat Daut, M. A., Hassan, M. Y., Abdullah, H., Rahman, H. A., Abdullah, M. P., Hussin, F., 2017. Building electrical energy consump-
tion forecasting analysis using conventional and artificial intelligence methods: A review. Renewable and Sustainable Energy
Reviews 70, 1108–1118.
Matte, T. D., Lane, K., Ito, K., 2016. Excess mortality attributable to extreme heat in New York City, 1997-2013. Health Security 14 (2),
64–70.
McAlinn, K., West, M., 2019. Dynamic Bayesian predictive synthesis in time series forecasting. Journal of Econometrics 210 (1),
155–169.
McCarthy, C., Ryan, T. M., 1977. Estimates of voter transition probabilities from the British general elections of 1974. Journal of the
Royal Statistical Society, Series A 140 (1), 78–85.
McCoy, T. H., Pellegrini, A. M., Perlis, R. H., 2018. Assessment of time-series machine learning methods for forecasting hospital
discharge volume. JAMA network open 1 (7), e184087–e184087.
McLean, R. D., Pontiff, J., 2016. Does academic research destroy return predictability? Journal of Finance 71 (1), 5–32.
McNees, S. K., 1990. The role of judgment in macroeconomic forecasting accuracy. International Journal of Forecasting 6 (3), 287–299.
McNeil, A. J., Frey, R., Embrechts, P., 2015. Quantitative Risk Management: Concepts, Techniques and Tools - Revised Edition.
Princeton University Press.
Meade, N., 1984. The use of growth curves in forecasting market development - a review and appraisal. Journal of Forecasting 3 (4),
429–451.
Meade, N., 2000. Evidence for the selection of forecasting methods. Journal of Forecasting 19 (6), 515–535.
Meade, N., Islam, T., 2006. Modelling and forecasting the diffusion of innovation – a 25-year review. International Journal of Fore-
casting 22, 519–545.
Meade, N., Islam, T., 2015a. Forecasting in telecommunications and ICT - a review. International Journal of Forecasting 31 (4), 1105–
1126.
Meade, N., Islam, T., 2015b. Modelling European usage of renewable energy technologies for electricity generation. Technological
Forecasting and Social Change 90, 497–509.
Medeiros, M. C., Teräsvirta, T., Rech, G., 2006. Building neural network models for time series: a statistical approach. Journal of
Forecasting 25 (1), 49–75.
Meehl, P., 2013. Clinical Versus Statistical Prediction: A Theoretical Analysis and a Review of the Evidence. Echo Point Books &
Media.
Meese, R. A., Rogoff, K., 1983. Empirical exchange rate models of the seventies: Do they fit out of sample? Journal of International
Economics 14 (1-2), 3–24.
Meinshausen, N., 2006. Quantile regression forests. Journal of Machine Learning Research 7, 983–999.
Meira, E., Cyrino Oliveira, F. L., Jeon, J., 2020. Treating and pruning: new approaches to forecasting model selection and combination
using prediction intervals. International Journal of Forecasting.
Melacini, M., Perotti, S., Rasini, M., Tappia, E., 2018. E-fulfilment and distribution in omni-channel retailing: a systematic literature
review. International Journal of Physical Distribution & Logistics Management 48 (4), 391–414.
Mellit, A., Massi Pavan, A., Ogliari, E., Leva, S., Lughi, V., 2020. Advanced methods for photovoltaic output power forecasting: A
review. Applied Sciences 10 (2), 487.
Mello, J., 2009. The impact of sales forecast game playing on supply chains. Foresight: The International Journal of Applied Forecast-
ing 13, 13–22.
Mello, J., 2010. Corporate culture and S&OP: Why culture counts. Foresight: The International Journal of Applied Forecasting 16,
46–49.
Mena-Oreja, J., Gozalvez, J., 2020. A comprehensive evaluation of deep learning-based techniques for traffic prediction. IEEE Access
8, 91188–91212.
Meng, X., Bradley, J., Yavuz, B., Sparks, E., Venkataraman, S., Liu, D., Freeman, J., Tsai, D., Amde, M., Owen, S., Xin, D., Xin, R.,
Franklin, M. J., Zadeh, R., Zaharia, M., Talwalkar, A., 2016. MLlib: Machine Learning in Apache Spark. The Journal of Machine

217
 Learning Research 17 (1), 1235–1241.
Meng, X., Taylor, J. W., 2020. Estimating Value-at-Risk and Expected Shortfall using the intraday low and range data. European
Journal of Operational Research 280 (1), 191–202.
Meng, X., Taylor, J. W., Ben Taieb, S., Li, S., 2020. Scoring functions for multivariate distributions and level sets. arXiv:2002.09578.
Merkle, E. C., Steyvers, M., 2013. Choosing a strictly proper scoring rule. Decision Analysis 10 (4), 292–304.
Merrick, J. R. W., Hardin, J. R., Walker, R., 2006. Partnerships in training. INFORMS Journal on Applied Analytics 36 (4), 359–370.
Merrow, E. W., McDonnell, L. M., Arguden, R. Y., 1988. Understanding the Outcomes of Mega-Projects. RAND Corporation.
Messner, J. W., Pinson, P., 2018. Online adaptive lasso estimation in vector autoregressive models for high dimensional wind power
forecasting. International Journal of Forecasting.
Mestre, G., Portela, J., San Roque, A. M., Alonso, E., 2020. Forecasting hourly supply curves in the italian day-ahead electricity market
with a double-seasonal SARMAHX model. International Journal of Electrical Power & Energy Systems 121, 106083.
Meyer, B. H., Tasci, M., 2015. Lessons for forecasting unemployment in the United States: Use flow rates, mind the trend. Working
paper no. 2015-1, FRB Atlanta.
Miao, D. W. C., Wu, C. C., Su, Y. K., 2013. Regime-switching in volatility and correlation structure using range-based models with
Markov-switching. Economic Modelling 31 (1), 87–93.
Miao, H., Ramchander, S., Wang, T., Yang, D., 2017. Influential factors in crude oil price forecasting. Energy Economics 68, 77–88.
Mikkelsen, L., Moesgaard, K., Hegnauer, M., Lopez, A. D., 2020. Anaconda: a new tool to improve mortality and cause of death data.
BMC Medicine 18 (1), 1–13.
Milankovitch, M., 1969. Canon of insolation and the ice-age problem. National Science Foundation, Washington, D.C, english trans-
lation by the Israel Program for Scientific Translations of Kanon der Erdbestrahlung und seine Anwendung auf das Eiszeitenproblem,
Textbook Publishing Company, Belgrade, 1941.
Milas, C., Rothman, P., 2008. Out-of-sample forecasting of unemployment rates with pooled STVECM forecasts. International Journal
of Forecasting 24 (1), 101–121.
Millán-Ruiz, D., Hidalgo, J. I., 2013. Forecasting call centre arrivals. Journal of Forecasting 32 (7), 628–638.
Miller, R., Lessard, D., 2007. Evolving strategy: Risk management and the shaping of large engineering projects. Tech. Rep. 37157,
Massachusetts Institute of Technology.
Min, A., Czado, C., 2011. Bayesian model selection for D-vine pair-copula constructions. Canadian Journal of Statistics 39 (2), 239–
258.
Min, C.-k., Zellner, A., 1993. Bayesian and non-Bayesian methods for combining models and forecasts with applications to forecasting
international growth rates. Journal of Econometrics 56 (1-2), 89–118.
Mincer, J., Zarnowitz, V., 1969. The evaluation of economic forecasts. In: Mincer, J. (Ed.), Economic Forecasts and Expectations:
Analysis of Forecasting Behavior and Performance. National Bureau of Economic Research, Inc, pp. 3–46.
Mingming, T., Jinliang, Z., 2012. A multiple adaptive wavelet recurrent neural network model to analyze crude oil prices. Journal of
Economics and Business 64 (4), 275–286.
Mirakyan, A., Meyer-Renschhausen, M., Koch, A., 2017. Composite forecasting approach, application for next-day electricity price
forecasting. Energy Economics 66, 228–237.
Mircetica, D., Rostami-Tabar, B., Nikolicica, S., Maslarica, M., 2020. Forecasting hierarchical time series in supply chains: an empirical
investigation, Cardiff University.
Mirko, K., Kantelhardt, J. W., 2013. Hadoop. TS: Large-scale time-series processing. International Journal of Computer Applications
74 (17).
Mirmirani, S., Li, H. C., 2004. A comparison of VAR and neural networks with genetic algorithm in forecasting price of oil. Advances
in Econometrics 19, 203–223.
Miron, J. A., 1996. The Economics of Seasonal Cycles. MIT Press.
Mišić, S., Radujković, M., 2015. Critical drivers of megaprojects success and failure. Procedia Engineering 122, 71–80.
Mitchell, T. J., Beauchamp, J. J., 1988. Bayesian variable selection in linear regression. Journal of the American Statistical Association
83 (404), 1023–1032.
Mitofsky, W., 1991. A short history of exit polls. In: Lavrakas, P. J., Holley, J. K. (Eds.), Polling and Presidential Election Coverage.
Sage, Newbury Park, CA, pp. 83–99.
Modis, T., 1992. Predictions: Society’s Telltale Signature Reveals the Past and Forecasts the Future. Simon & Schuster.
Modis, T., 1994. Fractal aspects of natural growth. Technological Forecasting and Social Change 47 (1), 63–73.
Modis, T., 1997. Genetic re-engineering of corporations. Technological Forecasting and Social Change 56 (2), 107–118.
Modis, T., 1998. Conquering Uncertainty: Understanding Corporate Cycles and Positioning Your Company to Survive the Changing

218
 Environment. McGraw-Hill.
Modis, T., 2013. Natural Laws in the Service of the Decision Maker: How to Use Science-Based Methodologies to See More Clearly
Further Into the Future. Growth Dynamics.
Modis, T., Debecker, A., 1992. Chaoslike states can be expected before and after logistic growth. Technological Forecasting and Social
Change 41 (2), 111–120.
Moghaddam, A. H., Moghaddam, M. H., Esfandyari, M., 2016. Stock market index prediction using artificial neural network. Journal
of Economics, Finance and Administrative Science 21 (41), 89–93.
Mohammadi, H., Su, L., 2010. International evidence on crude oil price dynamics: Applications of ARIMA-GARCH models. Energy
Economics 32 (5), 1001–1008.
Mohandes, S. R., Zhang, X., Mahdiyar, A., 2019. A comprehensive review on the application of artificial neural networks in building
energy analysis. Neurocomputing 340, 55–75.
Molenaers, A., Baets, H., Pintelon, L., Waeyenbergh, G., 2012. Criticality classification of spare parts: A case study. International
Journal of Production Economics 140 (2), 570–578.
Möller, A., Lenkoski, A., Thorarinsdottir, T. L., 2013. Multivariate probabilistic forecasting using ensemble Bayesian model averaging
and copulas. Quarterly Journal of the Royal Meteorological Society 139 (673), 982–991.
Molnár, P., 2016. High-low range in GARCH models of stock return volatility. Applied Economics 48 (51), 4977–4991.
Molodtsova, T., Papell, D. H., 2009. Out-of-sample exchange rate predictability with Taylor rule fundamentals. Journal of Interna-
tional Economics 77 (2), 167–180.
Monsell, B., Aston, J., Koopman, S., 2003. Toward X-13? In: Proceedings of the American Statistical Association, Section on Business
and Economic Statistics. pp. 1–8.
Montero Jimenez, J. J., Schwartz, S., Vingerhoeds, R., Grabot, B., Salaün, M., 2020. Towards multi-model approaches to predictive
maintenance: A systematic literature survey on diagnostics and prognostics. Journal of Manufacturing Systems 56, 539–557.
Montero-Manso, P., Athanasopoulos, G., Hyndman, R. J., Talagala, T. S., 2020. FFORMA: Feature-based forecast model averaging.
International Journal of Forecasting 36 (1), 86–92.
Montero-Manso, P., Hyndman, R. J., 2020. Principles and algorithms for forecasting groups of time series: Locality and globality.
arXiv:2008.00444.
Montgomery, A. L., Zarnowitz, V., Tsay, R. S., Tiao, G. C., 1998. Forecasting the U.S. unemployment rate. Journal of the American
Statistical Association 93, 478–493.
Moon, M. A., Mentzer, J. T., Smith, C. D., 2003. Conducting a sales forecasting audit. International Journal of Forecasting 19 (1), 5–25.
Moon, S., Simpson, A., Hicks, C., 2013. The development of a classification model for predicting the performance of forecasting
methods for naval spare parts demand. International Journal of Production Economics 143 (2), 449–454.
Mori, H., Yuihara, A., 2001. Deterministic annealing clustering for ANN-based short-term load forecasting. IEEE Transactions on
Power Systems 16 (3), 545–551.
Morlidge, S., 2014a. Do forecasting methods reduce avoidable error? Evidence from forecasting competitions. Foresight: The Inter-
national Journal of Applied Forecasting 32, 34–39.
Morlidge, S., 2014b. Forecast quality in the supply chain. Foresight: The International Journal of Applied Forecasting 33, 26–31.
Morlidge, S., 2014c. Using relative error metrics to improve forecast quality in the supply chain. Foresight: The International Journal
of Applied Forecasting 34, 39–46.
Morris, S., Pratt, D., 2003. Analysis of the Lotka–Volterra competition equations as a technological substitution model. Technological
Forecasting and Social Change 77, 103–133.
Morss, R. E., Demuth, J. L., Lazo, J. K., 2008. Communicating uncertainty in weather forecasts: A survey of the us public. Weather
and Forecasting 23 (5), 974–991.
Moshman, J., 1964. The role of computers in election night broadcasting. In: Alt, F. L., Rubinoff, M. (Eds.), Advances in Computers.
Vol. 5. Elsevier, pp. 1–21.
Mount, T. D., Ning, Y., Cai, X., 2006. Predicting price spikes in electricity markets using a regime-switching model with time-varying
parameters. Energy Economics 28 (1), 62–80.
Mueller, J., 2009a. Retreat from Doomsday: The Obsolescence of Major War. Zip Publishing.
Mueller, J., 2009b. War has almost ceased to exist: An assessment. Political Science Quarterly 124 (2), 297–321.
Muniain, P., Ziel, F., 2020. Probabilistic forecasting in day-ahead electricity markets: Simulating peak and off-peak prices. Interna-
tional Journal of Forecasting 36 (4), 1193–1210.
Murphy, A. H., 1993. What is a good forecast? An essay on the nature of goodness in weather forecasting. Weather and Forecasting
8 (2), 281–293.

219
Muth, J. F., 1961. Rational expectations and the theory of price movements. Econometrica 29, 315–335.
Nagi, J., Yap, K. S., Nagi, F., Tiong, S. K., Ahmed, S. K., 2011. A computational intelligence scheme for the prediction of the daily peak
load. Applied Soft Computing 11 (8), 4773–4788.
Nanopoulos, A., Alcock, R., Manolopoulos, Y., 2001. Feature-based classification of time-series data. In: Information processing and
technology. Nova Science Publishers, Inc., USA, pp. 49–61.
Narajewski, M., Ziel, F., 2020a. Econometric modelling and forecasting of intraday electricity prices. Journal of Commodity Markets
19, 100107.
Narajewski, M., Ziel, F., 2020b. Ensemble forecasting for intraday electricity prices: Simulating trajectories. Applied Energy 279,
115801.
National Research Council, 2006. Completing the forecast: Characterizing and communicating uncertainty for better decisions using
weather and climate forecasts. National Academies Press.
Neale, W. C., 1964. The Peculiar Economics of Professional Sports. Quarterly Journal of Economics 78 (1), 1–14.
Nelsen, R., 2006. An introduction to copulas. Springer Verlag.
Nelson, C. R., Plosser, C. R., 1982. Trends and random walks in macroeconomic time series: Some evidence and implications. Journal
of Monetary Economics 10 (2), 139–162.
Nelson, D. B., 1991. Conditional Heteroskedasticity in Asset Returns: A New Approach. Econometrica 59 (2), 347.
Nespoli, A., Ogliari, E., Leva, S., Massi Pavan, A., Mellit, A., Lughi, V., Dolara, A., 2019. Day-Ahead photovoltaic forecasting: A
comparison of the most effective techniques. Energies 12 (9), 1621.
Neves, M. M., Cordeiro, C., 2020. Modellling (and forecasting) extremes in time series: a naive approach. In: Atas do XXIII Congresso
da SPE. Sociedade Portuguesa de Estatı́stica, pp. 189–202.
Newbold, P., Granger, C. W., 1974. Experience with forecasting univariate time series and the combination of forecasts. Journal of the
Royal Statistical Society: Series A (General) 137 (2), 131–146.
Ng, Y. S., Stein, J., Ning, M., Black-Schaffer, R. M., 2007. Comparison of clinical characteristics and functional outcomes of ischemic
stroke in different vascular territories. Stroke 38 (8), 2309–2314.
Nielsen, J., Mazick, A., Andrews, N., Detsis, M., Fenech, T., Flores, V., Foulliet, A., Gergonne, B., Green, H., Junker, C., Nunes, B.,
O’Donnell, J., Oza, A., Paldy, A., Pebody, R., Reynolds, A., Sideroglou, T., E, S. B., Simon-Sofia, F., Uphoff, H., Van Asten, L.,
Virtanen, M. J., Wuillaume, F., Molbak, K., 2013. Pooling European all-cause mortality: methodology and findings for the seasons
2008/2009 to 2010/2011. Epidemiology & Infection 141 (9), 1996–2010.
Nielsen, M., Seo, W., Seong, D., 2019. Inference on the dimension of the nonstationary subspace in functional time series. Working
Paper 1420, Queen’s Economics Department.
Nikolopoulos, K., 2020. We need to talk about intermittent demand forecasting. European Journal of Operational Research.
Nikolopoulos, K., Assimakopoulos, V., Bougioukos, N., Litsa, A., Petropoulos, F., 2012. The theta model: An essential forecasting tool
for supply chain planning. In: Advances in Automation and Robotics, Vol. 2. Springer Berlin Heidelberg, pp. 431–437.
Nikolopoulos, K., Goodwin, P., Patelis, A., Assimakopoulos, V., 2007. Forecasting with cue information: A comparison of multiple
regression with alternative forecasting approaches. European Journal of Operational Research 180 (1), 354–368.
Nikolopoulos, K., Litsa, A., Petropoulos, F., Bougioukos, V., Khammash, M., 2015. Relative performance of methods for forecasting
special events. Journal of Business Research 68 (8), 1785–1791.
Nikolopoulos, K., Petropoulos, F., 2018. Forecasting for big data: Does suboptimality matter? Computers & Operations Research 98,
322–329.
Nikolopoulos, K., Punia, S., Schäfers, A., Tsinopoulos, C., Vasilakis, C., 2020. Forecasting and planning during a pandemic: COVID-
19 growth rates, supply chain disruptions, and governmental decisions. European Journal of Operational Research.
Nikolopoulos, K., Syntetos, A. A., Boylan, J. E., Petropoulos, F., Assimakopoulos, V., 2011. An aggregate - disaggregate intermittent
demand approach (ADIDA) to forecasting: An empirical proposition and analysis. The Journal of the Operational Research Society
62 (3), 544–554.
Nikolopoulos, K. I., Babai, M. Z., Bozos, K., 2016. Forecasting supply chain sporadic demand with nearest neighbor approaches.
International Journal of Production Economics 177, 139–148.
Nikolopoulos, K. I., Thomakos, D. D., 2019. Forecasting With The Theta Method: Theory and Applications. John Wiley & Sons.
Nogueira, P. J., de Araújo Nobre, M., Nicola, P. J., Furtado, C., Carneiro, A. V., 2020. Excess mortality estimation during the COVID-19
pandemic: preliminary data from Portugal. Acta Médica Portuguesa 33 (13).
Nordhaus, W. D., 1987. Forecasting Efficiency: Concepts and Applications. The Review of Economics and Statistics 69 (4), 667–674.
Norton-Taylor, R., 2015. Global armed conflicts becoming more deadly, major study finds. http://www.theguardian.com/world/
2015/may/20/, accessed on 2020-09-12.

220
Nowotarski, J., Weron, R., 2018. Recent advances in electricity price forecasting: A review of probabilistic forecasting. Renewable
and Sustainable Energy Reviews 81, 1548–1568.
Nsoesie, E., Mararthe, M., Brownstein, J., 2013. Forecasting peaks of seasonal influenza epidemics. PLoS Currents 5.
Nunes, B., Viboud, C., Machado, A., Ringholz, C., Rebelo-de Andrade, H., Nogueira, P., Miller, M., 2011. Excess mortality associated
with influenza epidemics in Portugal, 1980 to 2004. PloS One 6 (6), e20661.
Nye, J. S., 1990. The changing nature of world power. Political Science Quarterly 105 (2), 177–192.
Nymoen, R., Sparrman, V., 2015. Equilibrium unemployment dynamics in a panel of OECD countries. Oxford Bulletin of Economics
and Statistics 77 (2), 164–190.
Nystrup, P., Lindström, E., Pinson, P., Madsen, H., 2020. Temporal hierarchies with autocorrelation for load forecasting. European
Journal of Operational Research 280 (3), 876–888.
OBR, 2019. Long-term economic determinants. OBR Supplementary Forecast Information Release, Office for Budget Responsibility,
London.
Obst, D., Ghattas, B., Claudel, S., Cugliari, J., Goude, Y., Oppenheim, G., 2019. Textual data for time series forecasting.
arXiv:1910.12618.
O’Connor, M., Remus, W., Griggs, K., 1993. Judgemental forecasting in times of change. International Journal of Forecasting 9 (2),
163–172.
Office for National Statistics, 2019. U.K. national accounts, the blue book: 2019. Tech. rep., Office for National Statistics.
Ogata, Y., 1978. The asymptotic behaviour of maximum likelihood estimators for stationary point processes. Annals of the Institute
of Statistical Mathematics 30 (2), 243–261.
Ogata, Y., 1988. Statistical models for earthquake occurrences and residual analysis for point processes. Journal of the American
Statistical Association 83 (401), 9–27.
Ogliari, E., Dolara, A., Manzolini, G., Leva, S., 2017. Physical and hybrid methods comparison for the day ahead PV output power
forecast. Renewable Energy 113, 11–21.
Ogliari, E., Niccolai, A., Leva, S., Zich, R. E., 2018. Computational intelligence techniques applied to the day ahead PV output power
forecast: PHANN, SNO and mixed. Energies 11 (6), 1487.
Oh, D. H., Patton, A. J., 2016. High-dimensional copula-based distributions with mixed frequency data. Journal of Econometrics
193 (2), 349–366.
Oh, D. H., Patton, A. J., 2018. Time-varying systemic risk: Evidence from a dynamic copula model of cds spreads. Journal of Business
& Economic Statistics 36 (2), 181–195.
Oh, H., Yoon, C., 2020. Time to build and the real-options channel of residential investment. Journal of Financial Economics 135 (1),
255–269.
O’Hagan, A., Buck, C. E., Daneshkhah, A., Richard Eiser, J., Garthwaite, P. H., Jenkinson, D. J., Oakley, J. E., Rakow, T., 2006. Uncertain
Judgements: Eliciting Experts’ Probabilities. Wiley.
O’Hara-Wild, M., Hyndman, R., 2020. fasster: Fast Additive Switching of Seasonality, Trend and Exogenous Regressors. R package
version 0.1.0.9100.
Oksuz, I., Ugurlu, U., 2019. Neural network based model comparison for intraday electricity price forecasting. Energies 12 (23), 4557.
Okun, A. M., 1962. Potential GNP: Its measurement and significance. American Statistical Association, Proceedings of the Business
and Economics Statistics Section, 98–104.
Oliva, R., Watson, N., 2009. Managing functional biases in organizational forecasts: A case study of consensus forecasting in supply
chain planning. International Journal of Operations & Production Management 18 (2), 138–151.
Oliveira, F. L. C., Souza, R. C., Marcato, A. L. M., 2015. A time series model for building scenarios trees applied to stochastic optimi-
sation. International Journal of Electrical Power & Energy Systems 67, 315–323.
Oliveira, J. M., Ramos, P., 2019. Assessing the performance of hierarchical forecasting methods on the retail sector. Entropy 21 (4).
Önkal, D., Gönül, M. S., 2005. Judgmental adjustment: A challenge for providers and users of forecasts. Foresight: The International
Journal of Applied Forecasting 1, 13–17.
Önkal, D., Gönül, M. S., De Baets, S., 2019. Trusting forecasts. Futures & Foresight Science 1, e19.
Önkal, D., Gönül, M. S., Goodwin, P., Thomson, M., Öz, E., 2017. Evaluating expert advice in forecasting: Users’ reactions to presumed
vs. experienced credibility. International Journal of Forecasting 33 (1), 280–297.
Önkal, D., Gönül, M. S., Lawrence, M., 2008. Judgmental adjustments of previously adjusted forecasts. Decision Sciences 39 (2),
213–238.
Önkal, D., Goodwin, P., Thomson, M., Gönül, M. S., Pollock, A., 2009. The relative influence of advice from human experts and
statistical methods on forecast adjustments. Journal of Behavioral Decision Making 22 (4), 390–409.

221
Önkal, D., Sayım, K. Z., Gönül, M. S., 2013. Scenarios as channels of forecast advice. Technological Forecasting and Social Change
80 (4), 772–788.
Ord, J. K., Fildes, R., Kourentzes, N., 2017. Principles of Business Forecasting, 2nd Edition. Wessex Press Publishing Co.
Ord, K., Fildes, R., 2013. Principles of Business Forecasting, 1st Edition. South-Western Cengage Learning, Mason, OH and Andover,
UK.
Ordu, M., Demir, E., Tofallis, C., 2019. A comprehensive modelling framework to forecast the demand for all hospital services. The
International Journal of Health Planning and Management 34 (2), e1257–e1271.
Oreshkin, B. N., Carpov, D., Chapados, N., Bengio, Y., 2020a. Meta-learning framework with applications to zero-shot time-series
forecasting. arXiv:2002.02887.
Oreshkin, B. N., Carpov, D., Chapados, N., Bengio, Y., 2020b. N-BEATS: Neural basis expansion analysis for interpretable time series
forecasting. arXiv:1905.10437.
Ozaki, T., 1979. Maximum likelihood estimation of Hawkes’ self-exciting point processes. Annals of the Institute of Statistical Math-
ematics 31 (1), 145–155.
Özer, Ö., Zheng, Y., Chen, K.-Y., 2011. Trust in forecast information sharing. Management Science 57 (6), 1111–1137.
Paap, R., Franses, P. H., Hoek, H., 1997. Mean shifts, unit roots and forecasting seasonal time series. International Journal of Forecast-
ing 13 (3), 357–368.
Paccagnini, A., 2017. Dealing with Misspecification in DSGE Models: A Survey. MPRA Paper 82914, University Library of Munich,
Germany.
Pacheco, J., Millán-Ruiz, D., Vélez, J. L., 2009. Neural networks for forecasting in a multi-skill call centre. In: International Conference
on Engineering Applications of Neural Networks. Springer, pp. 291–300.
Pai, J., Pedersen, H., 1999. Threshold models of the term structure of interest rate. In: Joint day Proceedings Volume of the XXXth
International ASTIN Colloquium/9th International AFIR Colloquium. Tokyo, Japan, pp. 387–400.
Paillard, D., 2001. Glacial cycles: towards a new paradigm. Reviews of Geophysics 39, 325–346.
Pal, D., Mitra, S. K., 2019. Correlation dynamics of crude oil with agricultural commodities: A comparison between energy and food
crops. Economic Modelling 82, 453–466.
Palm, F. C., Zellner, A., 1992. To combine or not to combine? issues of combining forecasts. Journal of Forecasting 11 (8), 687–701.
Panagiotelis, A., Czado, C., Joe, H., 2012. Pair copula constructions for multivariate discrete data. Journal of the American Statistical
Association 107 (499), 1063–1072.
Panagiotelis, A., Czado, C., Joe, H., Stöber, J., 2017. Model selection for discrete regular vine copulas. Computational Statistics & Data
Analysis 106, 138–152.
Panagiotelis, A., Gamakumara, P., Athanasopoulos, G., Hyndman, R. J., 2020. Forecast reconciliation: A geometric view with new
insights on bias correction. International Journal of Forecasting in press.
Panahifar, F., Byrne, P. J., Heavey, C., 2015. A hybrid approach to the study of CPFR implementation enablers. Production Planning
& Control 26 (13), 1090–1109.
Panda, C., Narasimhan, V., 2007. Forecasting exchange rate better with artificial neural network. Journal of Policy Modeling 29 (2),
227–236.
Parag, Y., Sovacool, B. K., 2016. Electricity market design for the prosumer era. Nature Energy 1 (4), 1–6.
Paredes, J., Pedregal, D. J., Pérez, J. J., 2014. Fiscal policy analysis in the euro area: Expanding the toolkit. Journal of Policy Modeling
36, 800–823.
Park, B.-J., 2002. An outlier robust GARCH model and forecasting volatility of exchange rate returns. Journal of Forecasting 21 (5),
381–393.
Park, J., Sandberg, I. W., 1991. Universal approximation using Radial-Basis-Function networks. Neural Computation 3 (2), 246–257.
Park, S., Rabinovich, E., Tang, C. S., Yin, R., 2020. The impact of disclosing inventory-scarcity messages on sales in online retailing.
Journal of Operations Management 66 (5), 534–552.
Parkinson, M., 1980. The extreme value method for estimating the variance of the rate of return. The Journal of Business 53 (1),
61–65.
Pastore, E., Alfieri, A., Zotteri, G., Boylan, J. E., 2020. The impact of demand parameter uncertainty on the bullwhip effect. European
Journal of Operational Research 283 (1), 94–107.
Patel, J. K., 1989. Prediction intervals - a review. Communications in Statistics - Theory and Methods 18 (7), 2393–2465.
Patterson, K. D., 1995. An integrated model of the data measurement and data generation processes with an application to consumers’
expenditure. Economic Journal 105, 54–76.
Patti, E., Acquaviva, A., Jahn, M., Pramudianto, F., Tomasi, R., Rabourdin, D., Virgone, J., Macii, E., 2016. Event-Driven User-Centric

222
 Middleware for Energy-Efficient Buildings and Public Spaces. IEEE Systems Journal 10 (3), 1137–1146.
Patton, A., 2006a. Modelling asymmetric exchange rate dependence. International Economic Review 47 (2), 527–556.
Patton, A., 2013. Copula methods for forecasting multivariate time series. In: Handbook of economic forecasting. Vol. 2. Elsevier, pp.
899–960.
Patton, A. J., 2006b. Estimation of multivariate models for time series of possibly different lengths. Journal of Applied Econometrics
21 (2), 147–173.
Patton, A. J., Timmermann, A., 2007. Testing forecast optimality under unknown loss. Journal of the American Statistical Association
102, 1172–1184.
Pavı́a, J. M., 2010. Improving predictive accuracy of exit polls. International Journal of Forecasting 26 (1), 68–81.
Pavia, J. M., Cabrer, B., Sala, R., 2009. Updating input–output matrices: assessing alternatives through simulation. Journal of Statis-
tical Computation and Simulation 79 (12), 1467–1482.
Pavı́a, J. M., Gil-Carceller, I., Rubio-Mataix, A., Coll, V., Alvarez-Jareño, J. A., Aybar, C., Carrasco-Arroyo, S., 2019. The formation of
aggregate expectations: wisdom of the crowds or media influence? Contemporary Social Science 14 (1), 132–143.
Pavı́a, J. M., Larraz, B., 2012. Nonresponse bias and superpopulation models in electoral polls. Reis 137 (1), 237–264.
Pavı́a-Miralles, J. M., 2005. Forecasts from nonrandom samples. Journal of the American Statistical Association 100 (472), 1113–1122.
Pavı́a-Miralles, J. M., Larraz-Iribas, B., 2008. Quick counts from non-selected polling stations. Journal of Applied Statistics 35 (4),
383–405.
Payne, J. W., 1982. Contingent decision behavior. Psychological Bulletin 92 (2), 382–402.
Pearl, J., 2009. Causality: Models, Reasoning, and Inference, 2nd Edition. Cambridge University Press.
Pedregal, D. J., Carmen Carnero, M., 2006. State space models for condition monitoring: a case study. Reliability Engineering &
System Safety 91 (2), 171–180.
Pedregal, D. J., Garcı́a, F. P., Roberts, C., 2009. An algorithmic approach for maintenance management based on advanced state space
systems and harmonic regressions. Annals of Operations Research 166 (1), 109–124.
Pedregal, D. J., Pérez, J. J., 2010. Should quarterly government finance statistics be used for fiscal surveillance in Europe? International
Journal of Forecasting 26, 794–807.
Pedregal, D. J., Pérez, J. J., Sánchez, A. J., 2014. A toolkit to strengthen government budget surveillance. Review of Public Economics
211, 117–146.
Peel, D. A., Speight, A., 2000. Threshold nonlinearities in unemployment rates: Further evidence for the UK and G3 economies.
Applied Economics 32 (6), 705–715.
Pegels, C. C., 1969. Exponential forecasting: Some new variations. Management Sience 15 (5), 311–315.
Pelletier, D., 2006. Regime switching for dynamic correlations. Journal of Econometrics 131 (1), 445–473.
Pennings, C. L., van Dalen, J., 2017. Integrated hierarchical forecasting. European Journal of Operational Research 263 (2), 412–418.
Pennings, C. L. P., van Dalen, J., Rook, L., 2019. Coordinating judgmental forecasting: Coping with intentional biases. Omega 87,
46–56.
Perera, H. N., Hurley, J., Fahimnia, B., Reisi, M., 2019. The human factor in supply chain forecasting: A systematic review. European
Journal of Operational Research 274 (2), 574–600.
Peres, R., Muller, E., Mahajan, V., 2010. Innovation diffusion and new product growth models: A critical review and research direc-
tions. International Journal of Research in Marketing 27, 91–106.
Perron, P., 1989. The great crash, the oil price shock, and the unit root hypothesis. Econometrica 57 (6), 1361–1401.
Pesaran, M. H., Pick, A., Pranovich, M., 2013. Optimal forecasts in the presence of structural breaks. Journal of Econometrics 177 (2),
134–152.
Peters, J., Janzing, D., Schölkopf, B., 2017. Elements of Causal Inference. MIT Press.
Petropoulos, F., 2015. Forecasting support systems: Ways forward. Foresight: The International Journal of Applied Forecasting 39,
5–11.
Petropoulos, F., Fildes, R., Goodwin, P., 2016. Do ‘big losses’ in judgmental adjustments to statistical forecasts affect experts’ be-
haviour? European Journal of Operational Research 249 (3), 842–852.
Petropoulos, F., Goodwin, P., Fildes, R., 2017. Using a rolling training approach to improve judgmental extrapolations elicited from
forecasters with technical knowledge. International Journal of Forecasting 33 (1), 314–324.
Petropoulos, F., Hyndman, R. J., Bergmeir, C., 2018a. Exploring the sources of uncertainty: Why does bagging for time series fore-
casting work? European Journal of Operational Research 268 (2), 545–554.
Petropoulos, F., Kourentzes, N., 2014. Improving forecasting via multiple temporal aggregation. Foresight: The International Journal
of Applied Forecasting 34, 12–17.

223
Petropoulos, F., Kourentzes, N., 2015. Forecast combinations for intermittent demand. The Journal of the Operational Research
Society 66 (6), 914–924.
Petropoulos, F., Kourentzes, N., Nikolopoulos, K., Siemsen, E., 2018b. Judgmental selection of forecasting models. Journal of Opera-
tions Management 60, 34–46.
Petropoulos, F., Makridakis, S., 2020. Forecasting the novel coronavirus COVID-19. PloS One 15 (3), e0231236.
Petropoulos, F., Makridakis, S., Assimakopoulos, V., Nikolopoulos, K., 2014. ‘Horses for Courses’ in demand forecasting. European
Journal of Operational Research 237 (1), 152–163.
Petropoulos, F., Spyros, M., Stylianou, N., 2020. COVID-19: Forecasting confirmed cases and deaths with a simple time-series model.
International Journal of Forecasting.
Pfann, G. A., Schotman, P. C., Tschernig, R., 1996. Nonlinear interest rate dynamics and implications for the term structure. Journal
of Econometrics 74 (1), 149–176.
Phillips, A. W. H., 1958. The relation between unemployment and the rate of change of money wage rates in the United Kingdom,
1861–1957. Economica 25, 283–299.
Phillips, P. C. B., 1987. Time series regression with a unit root. Econometrica 55 (2), 277–301.
Phillips, R. F., 1996. Forecasting in the presence of large shocks. Journal of Economic Dynamics & Control 20 (9), 1581–1608.
Pierce, M. A., Hess, E. P., Kline, J. A., Shah, N. D., Breslin, M., Branda, M. E., Pencille, L. J., Asplin, B. R., Nestler, D. M., Sadosty, A. T.,
Stiell, I. G., Ting, H. H., Montori, V. M., 2010. The chest pain choice trial: a pilot randomized trial of a decision aid for patients
with chest pain in the emergency department. Trials 11 (1), 1–8.
Piironen, J., Vehtari, A., 2017. Comparison of Bayesian predictive methods for model selection. Statistics and Computing 27 (3),
711–735.
Pinheiro Neto, D., Domingues, E. G., Coimbra, A. P., de Almeida, A. T., Alves, A. J., Calixto, W. P., 2017. Portfolio optimization of
renewable energy assets: Hydro, wind, and photovoltaic energy in the regulated market in Brazil. Energy Economics 64, 238–250.
Pinker, S., 2011. The Better Angels of Our Nature: The Decline of Violence In History And Its Causes. Penguin UK.
Pinker, S., 2018. Enlightenment Now: The Case for Reason, Science, Humanism, and Progress. Penguin.
Pinson, P., 2012. Very-short-term probabilistic forecasting of wind power with generalized logit-normal distributions. Journal of the
Royal Statistical Society: Series C, 555–576.
Pinson, P., Chevallier, C., Kariniotakis, G., 2007. Trading wind generation from short-term probabilistic forecasts of wind power.
IEEE Transaction on Power Systems 22 (3), 1148–1156.
Pinson, P., Madsen, H., Nielsen, H. A., Papaefthymiou, G., Klöckl, B., 2009. From probabilistic forecasts to statistical scenarios of
short-term wind power production. Wind Energy 12 (1), 51–62.
Pinson, P., Makridakis, S., 2020. Pandemics and forecasting: The way forward through the Taleb-Ioannidis debate. International
Journal of Forecasting.
Pinson, P., Reikard, G., Bidlot, J.-R., 2012. Probabilistic forecasting of the wave energy flux. Applied Energy 93, 364–370.
Pinson, P., Tastu, J., 2013. Discrimination ability of the energy score. Technical University of Denmark (DTU).
Pirolli, P., Card, S., 1999. Information foraging. Psychological Review 106 (4), 643–675.
Pitt, M., Chan, D., Kohn, R., 2006. Efficient Bayesian inference for Gaussian copula regression models. Biometrika 93 (3), 537–554.
Plescia, C., De Sio, L., 2018. An evaluation of the performance and suitability of R × C methods for ecological inference with known
true values. Quality & Quantity 52 (2), 669–683.
Poccia, D., 2019. Amazon forecast – now generally available. https://aws.amazon.com/blogs/aws/
amazon-forecast-now-generally-available/, accessed on 2020-09-01.
Politi, M. C., Han, P. K., Col, N. F., 2007. Communicating the uncertainty of harms and benefits of medical interventions. Medical
Decision Making 27 (5), 681–695.
Politis, D. N., Romano, J. P., 1992. A circular block-resampling procedure for stationary data. Exploring the Limits of Bootstrap
2635270.
Politis, D. N., Romano, J. P., 1994. The stationary bootstrap. Journal of the American Statistical Association 89 (428), 1303–1313.
Polk, C., Haghbin, M., de Longis, A., 2020. Time-series variation in factor premia: The influence of the business cycle. Journal of
Investment Management 18 (1).
Polson, N. G., Sokolov, V. O., 2017. Deep learning for short-term traffic flow prediction. Transportation Research Part C: Emerging
Technologies 79, 1–17.
Porras, E., Dekker, R., 2008. An inventory control system for spare parts at a refinery: An empirical comparison of different re-order
point methods. European Journal of Operational Research 184 (1), 101–132.
Powell, W. B., 2019. A unified framework for stochastic optimization. European Journal of Operational Research 275 (3), 795–821.

224
Poynting, J. H., 1884. A comparison of the fluctuations in the price of wheat and in the cotton and silk imports into Great Britain.
Journal of the Statistical Society of London 47 (1), 34–74.
Pradeepkumar, D., Ravi, V., 2017. Forecasting financial time series volatility using particle swarm optimization trained quantile
regression neural network. Applied Soft Computing 58, 35–52.
Prahl, A., Van Swol, L., 2017. Understanding algorithm aversion: When is advice from automation discounted? Journal of Forecasting
36 (6), 691–702.
Prak, D., Teunter, R., Syntetos, A., 2017. On the calculation of safety stocks when demand is forecasted. European Journal of Opera-
tional Research 256 (2), 454–461.
Prestwich, S. D., Tarim, S. A., Rossi, R., Hnich, B., 2014. Forecasting intermittent demand by hyperbolic-exponential smoothing.
International Journal of Forecasting 30 (4), 928–933.
Pretis, F., 2020. Econometric modelling of climate systems: The equivalence of energy balance models and cointegrated vector au-
toregressions. Journal of Econometrics 214 (1), 256–273.
Pretis, F., Kaufmann, R. K., 2018. Out-of-sample Paleo-climate simulations: Testing hypotheses about the Mid-Brunhes event, the
Stage 11 paradox, and orbital variations. Discussion paper, University of Victoria, Canada.
Pretis, F., Kaufmann, R. K., 2020. Managing carbon emissions to avoid the next Ice Age. Discussion paper, University of Victoria,
Canada.
Pretis, F., Reade, J., Sucarrat, G., 2017. gets: General-to-Specific (GETS) Modelling and Indicator Saturation Methods. R package
version 0.12.
Pretis, F., Reade, J., Sucarrat, G., 2018. Automated General-to-Specific (GETS) Regression Modeling and Indicator Saturation for
Outliers and Structural Breaks. Journal of Statistical Software 86 (3).
Pretis, F., Schneider, L., Smerdon, J., 2016. Detecting volcanic eruptions in temperature reconstructions by designed break-indicator
saturation. Journal of Economic Surveys 30 (3), 403–429.
Proietti, T., 2003. Forecasting the US unemployment rate. Computational Statistics & Data Analysis 42, 451–476.
Promprou, S., Jaroensutasinee, M., Jaroensutasinee, K., 2006. Forecasting dengue haemorrhagic fever cases in southern Thailand
using ARIMA models. Dengue Bulletin 30, 99–106.
Prudêncio, R. B., Ludermir, T. B., 2004. Meta-learning approaches to selecting time series models. Neurocomputing 61, 121–137.
Psaradellis, I., Sermpinis, G., 2016. Modelling and trading the U.S. implied volatility indices. Evidence from the VIX, VXN and VXD
indices. International Journal of Forecasting 32 (4), 1268–1283.
Puig, X., Ginebra, J., 2015. Ecological inference and spatial variation of individual behavior: National divide and elections in catalo-
nia. Geographical Analysis 47 (3), 262–283.
Qiao, Z., Wu, X., Ge, S., Fan, W., 2019. MNN: multimodal attentional neural networks for diagnosis prediction. Extraction 1, A1.
Qu, X., Kang, X., Zhang, C., Jiang, S., Ma, X., 2016. Short-term prediction of wind power based on deep long Short-Term memory. In:
2016 IEEE PES Asia-Pacific Power and Energy Engineering Conference (APPEEC). pp. 1148–1152.
Quaedvlieg, R., 2019. Multi-horizon forecast comparison. Journal of Business & Economic Statistics, 1–14.
R Core Team, 2020. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna,
Austria.
Rabanser, S., Januschowski, T., Flunkert, V., Salinas, D., Gasthaus, J., 2020. The effectiveness of discretization in forecasting: An
empirical study on neural time series models. arXiv:2005.10111.
Racine, J., 2000. Consistent cross-validatory model-selection for dependent data: hv-block cross-validation. Journal of Econometrics
99 (1), 39–61.
Raftery, A. E., 1993. Bayesian model selection in structural equation models. In: Bollen, K. A., Long, J. S. (Eds.), Testing Structural
Equation Models. Sage, Newbury Park, CA, pp. 163–180.
Raftery, A. E., 2016. Use and communication of probabilistic forecasts. Statistical Analysis and Data Mining: The ASA Data Science
Journal 9 (6), 397–410.
Raftery, A. E., Madigan, D., Hoeting, J. A., 1997. Bayesian model averaging for linear regression models. Journal of the American
Statistical Association 92 (437), 179–191.
Rahman, S., Serletis, A., 2012. Oil price uncertainty and the Canadian economy: Evidence from a VARMA, GARCH-in-Mean, asym-
metric BEKK model. Energy Economics 34 (2), 603–610.
Rajvanshi, V., 2015. Performance of range and return based volatility estimators: evidence from Indian crude oil futures market.
Global Economy and Finance Journal 8 (1), 46–66.
Ramos, M., Mathevet, T., Thielen, J., Pappenberger, F., 2010. Communicating uncertainty in hydro-meteorological forecasts: mission
impossible? Meteorological Applications 17 (2), 223–235.

225
Ramos, P., Oliveira, J. M., 2016. A procedure for identification of appropriate state space and ARIMA models based on time-series
cross-validation. Algorithms 9 (4), 76.
Ramos, P., Santos, N., Rebelo, R., 2015. Performance of state space and ARIMA models for consumer retail sales forecasting. Robotics
and Computer-Integrated Manufacturing 34, 151–163.
Ranawana, R., Palade, V., 2006. Optimized precision-a new measure for classifier performance evaluation. In: 2006 IEEE International
Conference on Evolutionary Computation. IEEE, pp. 2254–2261.
Rangapuram, S. S., de Bezenac, E., Benidis, K., Stella, L., Januschowski, T., 2020. Normalizing Kalman filters for multivariate time
series analysis. In: Advances in Neural Information Processing Systems. pp. 7785–7794.
Rangapuram, S. S., Seeger, M. W., Gasthaus, J., Stella, L., Wang, Y., Januschowski, T., 2018. Deep state space models for time series
forecasting. In: Advances in Neural Information Processing Systems. pp. 7785–7794.
Rangarajan, P., Mody, S. K., Marathe, M., 2019. Forecasting dengue and influenza incidences using a sparse representation of Google
trends, electronic health records, and time series data. PLoS Computational Biology 15 (11), e1007518.
Ranjan, R., Gneiting, T., 2010. Combining probability forecasts. Journal of the Royal Statistical Society, Series B (Statistical Method-
ology) 72 (1), 71–91.
Rao, J. K., Anderson, L. A., Sukumar, B., Beauchesne, D. A., Stein, T., Frankel, R. M., 2010. Engaging communication experts in
a Delphi process to identify patient behaviors that could enhance communication in medical encounters. BMC Health Services
Research 10, 97.
Rao, K., Kishore, V., 2010. A review of technology diffusion models with special reference to renewable energy technologies. Renew-
able and Sustainable Energy Reviews 14 (3), 1070–1078.
Rapach, D. E., Strauss, J. K., 2009. Differences in housing price forecastability across U.S. states. International Journal of Forecasting
25 (2), 351–372.
Rapach, D. E., Strauss, J. K., Tu, J., Zhou, G., 2019. Industry return predictability: A machine learning approach. The Journal of
Financial Data Science 1 (3), 9–28.
Rapach, D. E., Strauss, J. K., Zhou, G., 2010. Out-of-sample equity premium prediction: Combination forecasts and links to the real
economy. Review of Financial Studies 23 (2), 821–862.
Rapach, D. E., Strauss, J. K., Zhou, G., 2013. International stock return predictability: What is the role of the United States? Journal
of Finance 68 (4), 1633–1662.
Rapach, D. E., Zhou, G., 2020. Time-series and cross-sectional stock return forecasting: New machine learning methods. In: Jur-
czenko, E. (Ed.), Machine Learning for Asset Management: New Developments and Financial Applications. Wiley, Hoboken, NJ,
pp. 1–34.
Rasp, S., Pritchard, M. S., Gentine, P., 2018. Deep learning to represent subgrid processes in climate models. Proceedings of the
National Academy of Sciences 115 (39), 9684–9689.
Ravishanker, N., Wu, L. S., Glaz, J., 1991. Multiple prediction intervals for time series: comparison of simultaneous and marginal
intervals. Journal of Forecasting 10 (5), 445–463.
Reade, J., Singleton, C., Brown, A., 2020. Evaluating Strange Forecasts: The Curious Case of Football Match Scorelines. Scottish
Journal of Political Economy.
Rebollo, J., Balakrishnan, H., 2014. Characterization and prediction of air traffic delays. Transportation research part C: Emerging
technologies 44, 231–241.
Reggiani, P., Boyko, O., 2019. A Bayesian processor of uncertainty for precipitation forecasting using multiple predictors and censor-
ing. Monthly Weather Review 147 (12), 4367–4387.
Reid, D., 1972. A comparison of forecasting techniques on economic time series. Forecasting in Action. Operational Research Society
and the Society for Long Range Planning.
Reikard, G., Pinson, P., Bidlot, J.-R., 2011. Forecasting ocean wave energy: The ECMWF wave model and time series methods. Ocean
Engineering 38 (10), 1089–1099.
Reikard, G., Robertson, B., Buckham, B., Bidlot, J.-R., Hiles, C., 2015. Simulating and forecasting ocean wave energy in western
Canada. Ocean Engineering 103, 223–236.
Reimers, S., Harvey, N., 2011. Sensitivity to autocorrelation in judgmental time series forecasting. International Journal of Forecasting
27 (4), 1196–1214.
Renzl, B., 2008. Trust in management and knowledge sharing: The mediating effects of fear and knowledge documentation. Omega
36 (2), 206–220.
Riahi, N., Hosseini-Motlagh, S.-M., Teimourpour, B., 2013. A three-phase hybrid times series modeling framework for improved
hospital inventory demand forecast. International Journal of Hospital Research 2 (3), 133–142.

226
Rice, G., Wirjanto, T., Zhao, Y., 2020. Tests for conditional heteroscedasticity of functional data. Journal of Time Series Analysis 41 (6),
733–758.
Richardson, L. F., 1948. Variation of the frequency of fatal quarrels with magnitude. Journal of the American Statistical Association
43 (244), 523–546.
Richardson, L. F., 1960. Statistics of Deadly Quarrels. Boxwood Press.
Rios, I., Wets, R. J.-B., Woodruff, D. L., 2015. Multi-period forecasting and scenario generation with limited data. Computational
Management Science 12, 267–295.
Ritchie, H., Ortiz-Ospina, E., Beltekian, D., Mathieu, E., Hasell, J., Macdonald, B., Giattino, C., Roser, M., 2020. Coronavirus Pandemic
(COVID-19). https://ourworldindata.org/coronavirus, accessed on 2020-08-20.
Riveiro, M., Helldin, T., Falkman, G., Lebram, M., 2014. Effects of visualizing uncertainty on decision-making in a target identification
scenario. Computers & Graphics 41, 84–98.
Roberts, J. M., 2001. Estimates of the productivity trend using time-varying parameter techniques. The BE Journal of Macroeconomics
1 (1).
Robinson, W. S., 1950. Ecological correlations and the behavior of individuals. American Sociological Review 15 (3), 351–357.
Rodriguez, J. C., 2007. Measuring financial contagion: A copula approach. Journal of Empirical Finance 14 (3), 401–423.
Rodrı́guez-Sanz, Á., Comendador, F., Valdés, R., Pérez-Castán, J., Montes, R. B., Serrano, S., 2019. Assessment of airport arrival
congestion and delay: Prediction and reliability. Transportation Research Part C: Emerging Technologies 98, 255–283.
Rogers, L. C. G., Satchell, S. E., 1991. Estimating variance from high, low and closing prices. The Annals of Applied Probability 1 (4),
504–512.
Romero, R., Pavı́a, J. M., Martı́n, J., Romero, G., 2020. Assessing uncertainty of voter transitions estimated from aggregated data.
application to the 2017 French presidential election. Journal of Applied Statistics 47 (13-15), 2711—-2736.
Rosen, O., Jiang, W., King, G., Tanner, M. A., 2001. Bayesian and frequentist inference for ecological inference: The RxC case. Statistica
Neerlandica 55 (2), 134–156.
Rosenblatt, M., 1952. Remarks on a multivariate transformation. Annals of Mathematical Statistics 23, 470–472.
Rossi, B., 2005. Testing long-horizon predictive ability with high persistence, and the meese–rogoff puzzle. International Economic
Review 46 (1), 61–92.
Rossi, B., 2013. Exchange rate predictability. Journal of Economic Literature 51 (4), 1063–1119.
Rossi, B., Sekhposyan, T., 2016. Forecast Rationality Tests in the Presence of Instabilities, with Applications to Federal Reserve and
Survey Forecasts. Journal of Applied Econometrics 31 (3), 507–532.
Rostami-Tabar, B., Babai, M. Z., Syntetos, A., Ducq, Y., 2013. Demand forecasting by temporal aggregation. Naval Research Logistics
60 (6), 479–498.
Rostami-Tabar, B., Ziel, F., 2020. Anticipating special events in emergency department forecasting. International Journal of Forecast-
ing.
Rothman, P., 1998. Forecasting asymmetric unemployment rates. Review of Economics and Statistics 80 (1), 164–168.
Rothschild, D., 2009. Forecasting elections: Comparing prediction markets, polls, and their biases. Public Opinion Quarterly 73 (5),
895–916.
Rottenberg, S., 1956. The Baseball Players’ Labor Market. The Journal of Political Economy 64 (3), 242–258.
Roulin, E., Vannitsem, S., 2019. Post-processing of seasonal predictions – case studies using EUROSIP hindcast data base. Nonlinear
Processes in Geophysics.
Rousseau, D. M., Sitkin, S. B., Burt, R. S., Camerer, C. F., 1998. Not so different after all: A cross-discipline view of trust. Academy of
Management Review 23 (3), 393–404.
Rowe, G., Wright, G., 2001. Expert opinions in forecasting: The role of the Delphi technique. In: Armstrong, J. S. (Ed.), Principles of
Forecasting: A Handbook for Researchers and Practitioners. Springer US, Boston, MA, pp. 125–144.
Royer, J. F., 1993. Review of recent advances in dynamical extended range forecasting for the extratropics. In: Shukla, J. (Ed.),
Prediction of Interannual Climate Variations. Springer, Berlin, Heidelberg, pp. 49–69.
Ruano, A. E., Crispim, E. M., Conceiçao, E. Z., Lúcio, M. M. J., 2006. Prediction of building’s temperature using neural networks
models. Energy and Buildings 38 (6), 682–694.
Rubaszek, M., 2020. Forecasting crude oil prices with DSGE models. International Journal of Forecasting.
Ruddiman, W., 2005. Plows, Plagues and Petroleum: How Humans took Control of Climate. Princeton University Press, Princeton.
Rycroft, R. S., 1993. Microcomputer software of interest to forecasters in comparative review: An update. International Journal of
Forecasting 9 (4), 531–575.
Sah, S., Moore, D. A., MacCoun, R. J., 2013. Cheap talk and credibility: The consequences of confidence and accuracy on advisor

227
 credibility and persuasiveness. Organizational Behavior and Human Decision Processes 121 (2), 246–255.
Sakamoto, Y., Ishiguro, M., Kitagawa, G., 1986. Akaike information criterion statistics. Dordrecht, The Netherlands: D. Reidel 81.
Sakata, S., White, H., 1998. High breakdown point conditional dispersion estimation with application to S&P 500 daily returns
volatility. Econometrica 66 (3), 529–568.
Saksornchai, T., Wei-Jen Lee, Methaprayoon, K., Liao, J. R., Ross, R. J., 2005. Improve the unit commitment scheduling by using the
neural-network-based short-term load forecasting. IEEE Transactions on Industry Applications 41 (1), 169–179.
Salinas, D., Bohlke-Schneider, M., Callot, L., Medico, R., Gasthaus, J., 2019a. High-dimensional multivariate forecasting with low-
rank gaussian copula processes. In: Advances in Neural Information Processing Systems. pp. 6827–6837.
Salinas, D., Flunkert, V., Gasthaus, J., Januschowski, T., 2019b. DeepAR: Probabilistic forecasting with autoregressive recurrent
networks. International Journal of Forecasting.
Salway, R., Wakefield, J., 2004. A common framework for ecological inference in epidemiology, political science and sociology. In:
Ecological Inference: New Methodological Strategies. Cambridge University Press, pp. 303–332.
Sanders, N. R., Manrodt, K. B., 2003. Forecasting software in practice: Use, satisfaction, and performance. Interfaces 33 (5), 90–93.
Sanderson, J., 2012. Risk, uncertainty and governance in megaprojects: A critical discussion of alternative explanations. International
Journal of Project Management 30 (4), 432–443.
Santos, M. S., Abreu, P. H., Garca-Laencina, P. J., Simão, A., Carvalho, A., 2015. A new cluster-based oversampling method for
improving survival prediction of hepatocellular carcinoma patients. Journal of Biomedical Informatics 58, 49–59.
Sardinha-Lourenço, A., Andrade-Campos, A., Antunes, A., Oliveira, M. S., 2018. Increased performance in the short-term water
demand forecasting through the use of a parallel adaptive weighting strategy. Journal of Hydrology 558, 392–404.
Savin, S., Terwiesch, C., 2005. Optimal product launch times in a duopoly: balancing life-cycle revenues with product cost. Operations
Research 53, 26–47.
Scerri, M., De Goumoens, P., Fritsch, C., Van Melle, G., Stiefel, F., So, A., 2006. The INTERMED questionnaire for predicting return
to work after a multidisciplinary rehabilitation program for chronic low back pain. Joint Bone Spine 73 (6), 736–741.
Schäfer, A. M., Zimmermann, H. G., 2006. Recurrent neural networks are universal approximators. In: Artificial Neural Networks –
ICANN 2006. Springer Berlin Heidelberg, pp. 632–640.
Scharpf, A., Schneider, G., Nöh, A., Clauset, A., 2014. Forecasting the risk of extreme massacres in Syria. European Review of Inter-
national Studies 1 (2), 50–68.
Schefzik, R., Thorarinsdottir, T. L., Gneiting, T., 2013. Uncertainty quantification in complex simulation models using ensemble
copula coupling. Statistical Science 28 (4), 616–640.
Scheuerer, M., Hamill, T. M., 2015. Variogram-based proper scoring rules for probabilistic forecasts of multivariate quantities.
Monthly Weather Review 143 (4), 1321–1334.
Scheuren, F. J., Alvey, W., 2008. Elections and Exit Polling. John Wiley & Sons, Hoboken (New Yersey).
Schnaars, S. P., Topol, M. T., 1987. The use of multiple scenarios in sales forecasting: An empirical test. International Journal of
Forecasting 3 (3), 405–419.
Schoemaker, P. J. H., 1991. When and how to use scenario planning: A heuristic approach with illustration. Journal of Forecasting
10 (6), 549–564.
Schönbucher, P. J., 2003. Credit derivatives pricing models: models, pricing and implementation. John Wiley & Sons.
Schubert, S., Rickard, R., 2011. Using forecast value added analysis for data-driven forecasting improvement. IBF Best Practices
Conference.
Schwanenberg, D., Fan, F. M., Naumann, S., Kuwajima, J. I., Montero, R. A., Assis dos Reis, A., 2015. Short-Term reservoir optimization
for flood mitigation under meteorological and hydrological forecast uncertainty. Water Resources Management 29 (5), 1635–1651.
Schwarz, G., 1978. Estimating the dimension of a model. Annals of Statistics 6 (2), 461–464.
Scott Armstrong, J., 2006. Should the forecasting process eliminate Face-to-Face meetings? Foresight: The International Journal of
Applied Forecasting 5, 3–8.
Seaman, B., 2018. Considerations of a retail forecasting practitioner. International Journal of Forecasting 34 (4), 822–829.
Seifert, D., 2003. Collaborative Planning, Forecasting, and Replenishment: How to Create a Supply Chain Advantage. AMACOM,
New York.
Semero, Y. K., Zhang, J., Zheng, D., 2020. EMD–PSO–ANFIS-based hybrid approach for short-term load forecasting in microgrids.
IET Generation, Transmission and Distribution 14 (3), 470–475.
Seong, Y., Bisantz, A. M., 2008. The impact of cognitive feedback on judgment performance and trust with decision aids. International
Journal of Industrial Ergonomics 38 (7), 608–625.
Serletis, A., Rangel-Ruiz, R., 2004. Testing for common features in North American energy markets. Energy Economics 26 (3), 401–

228
 414.
Setel, P., AbouZahr, C., Atuheire, E. B., Bratschi, M., Cercone, E., Chinganya, O., Clapham, B., Clark, S. J., Congdon, C., de Savigny,
D., Karpati, A., Nichols, E., Jakob, R., Mwanza, J., Muhwava, W., Nahmias, P., Ortiza, E. M., Tshangelab, A., 2020. Mortality
surveillance during the COVID-19 pandemic. Bulletin of the World Health Organization 98 (6), 374.
Setzler, H., Saydam, C., Park, S., 2009. EMS call volume predictions: A comparative study. Computers & Operations Research 36 (6),
1843–1851.
Shackleton, M. B., Taylor, S. J., Yu, P., 2010. A multi-horizon comparison of density forecasts for the S&P 500 using index returns and
option prices. Journal of Banking and Finance 34 (11), 2678–2693.
Shah, I., Lisi, F., 2020. Forecasting of electricity price through a functional prediction of sale and purchase curves. Journal of Fore-
casting 39 (2), 242–259.
Shahriari, M., Blumsack, S., 2018. The capacity value of optimal wind and solar portfolios. Energy 148, 992–1005.
Shale, E. A., Boylan, J. E., Johnston, F. R., 2006. Forecasting for intermittent demand: the estimation of an unbiased average. The
Journal of the Operational Research Society 57 (5), 588–592.
Shaman, J., Karspeck, A., 2012. Forecasting seasonal outbreaks of influenza. Proceedings of the National Academy of Sciences
109 (50), 20425–20430.
Shang, G., McKie, E. C., Ferguson, M. E., Galbreth, M. R., 2020. Using transactions data to improve consumer returns forecasting.
Journal of Operations Management 66 (3), 326–348.
Shang, H. L., Booth, H., Hyndman, R. J., 2011. Point and interval forecasts of mortality rates and life expectancy: A comparison of
ten principal component methods. Demographic Research 25, 173–214.
Shang, H. L., Haberman, S., 2020a. Forecasting age distribution of death counts: An application to annuity pricing. Annals of Actu-
arial Science 14 (1), 150–169.
Shang, H. L., Haberman, S., 2020b. Retiree mortality forecasting: A partial age-range or a full age-range model? Risks 8 (3), 69.
Shang, H. L., Hyndman, R. J., 2017. Grouped functional time series forecasting: An application to age-specific mortality rates. Journal
of Computational and Graphical Statistics 26 (2), 330–343.
Shang, J., Ma, T., Xiao, C., Sun, J., 2019. Pre-training of graph augmented transformers for medication recommendation.
arXiv:1906.00346.
Shapiro, A. F., 2000. A hitchhiker’s guide to the techniques of adaptive nonlinear models. Insurance, Mathematics & Economics 26 (2),
119–132.
Sharpe, W. F., 1964. Capital Asset Prices: A Theory of Market Equilibrium under Conditions of Risk. The Journal of Finance 19 (3),
425.
Shen, G., Jia, J., Nie, L., Feng, F., Zhang, C., Hu, T., Chua, T.-S., Zhu, W., 2017. Depression detection via harvesting social media: A
multimodal dictionary learning solution. In: IJCAI. pp. 3838–3844.
Shen, H., 2009. On modeling and forecasting time series of smooth curves. Technometrics 51 (3), 227–238.
Shen, H., Huang, J. Z., 2005. Analysis of call centre arrival data using singular value decomposition. Applied Stochastic Models in
Business and Industry 21 (3), 251–263.
Shen, H., Huang, J. Z., 2008a. Forecasting time series of inhomogeneous Poisson processes with application to call center workforce
management. The Annals of Applied Statistics 2 (2), 601–623.
Shen, H., Huang, J. Z., 2008b. Interday forecasting and intraday updating of call center arrivals. Manufacturing & Service Operations
Management 10 (3), 391–410.
Shen, H., Huang, J. Z., Lee, C., 2007. Forecasting and dynamic updating of uncertain arrival rates to a call center. In: 2007 IEEE
International Conference on Service Operations and Logistics, and Informatics. IEEE, pp. 1–6.
Sheng, C., Zhao, J., Leung, H., Wang, W., 2013. Extended Kalman Filter Based Echo State Network for Time Series Prediction using
MapReduce Framework. In: 2013 IEEE 9th International Conference on Mobile Ad-hoc and Sensor Networks. IEEE, pp. 175–180.
Shephard, N., 1994. Partial non-Gaussian state space. Biometrika 81, 115–131.
Shi, Q., Yin, J., Cai, J., Cichocki, A., Yokota, T., Chen, L., Yuan, M., Zeng, J., 2020. Block Hankel Tensor ARIMA for Multiple Short
Time Series Forecasting. In: AAAI. pp. 5758–5766.
Shishkin, J., Young, A. H., Musgrave, J. C., 1967. The X-11 variant of the Census II method seasonal adjustment program. Tech.
Rep. 15, Bureau of the Census, US Department of Commerce.
Shumway, R. H., Stoffer, D. S., 2017. Time Series Analysis and Its Applications: With R Examples. Springer.
Shvachko, K., Kuang, H., Radia, S., Chansler, R., 2010. The Hadoop Distributed File System. In: 2010 IEEE 26th Symposium on Mass
Storage Systems and Technologies (MSST). IEEE, pp. 1–10.
Si, X.-S., Wang, W., Hu, C.-H., Zhou, D.-H., 2011. Remaining useful life estimation – a review on the statistical data driven approaches.

229
 European Journal of Operational Research 213 (1), 1–14.
Sideratos, G., Ikonomopoulos, A., Hatziargyriou, N. D., 2020. A novel fuzzy-based ensemble model for load forecasting using hybrid
deep neural networks. Electric Power Systems Research 178, 106025.
Silvapulle, P., Moosa, I. A., 1999. The relationship between spot and futures prices: evidence from the crude oil market. Journal of
Futures Markets: Futures, Options, and Other Derivative Products 19 (2), 175–193.
Simpson, E. H., 1951. The interpretation of interaction in contingency tables. Journal of the Royal Statistical Society, Series B (Statis-
tical Methodology) 13 (2), 238–241.
Sims, C., 1980. Macroeconomics and Reality. Econometrica 48 (1), 1–48.
Sims, C., 2002. Solving linear rational expectations models. Computational Economics 20 (1-2), 1–20.
Sims, C. A., 2003. Implications of rational inattention. Journal of Monetary Economics 50, 665–690.
Singer, P. W., Friedman, A., 2014. Cybersecurity: What everyone needs to know. OUP USA.
Singh, P., Dwivedi, P., 2019. A novel hybrid model based on neural network and multi-objective optimization for effective load
forecast. Energy 182, 606–622.
Singleton, C., Reade, J., Brown, A., 2019. Going with your Gut: The (In)accuracy of Forecast Revisions in a Football Score Prediction
Game. Journal of Behavioral and Experimental Economics, 101502.
Sinnathamby, M. A., Whitaker, H., Coughlan, L., Bernal, J. L., Ramsay, M., Andrews, N., 2020. All-cause excess mortality observed by
age group and regions in the first wave of the COVID-19 pandemic in England. Eurosurveillance 25 (28), 2001239.
Smets, F., Warne, A., Wouters, R., 2014. Professional forecasters and real-time forecasting with a DSGE model. International Journal
of Forecasting 30 (4), 981–995.
Smets, F., Wouters, R., 2007. Shocks and Frictions in US Business Cycles: A Bayesian DSGE Approach. American Economic Review
97 (3), 586–606.
Smith, D. M., Scaife, A. A., Eade, R., Athanasiadis, P., Bellucci, A., Bethke, I., Bilbao, R., Borchert, L. F., Caron, L.-P., Counillon, F.,
Danabasoglu, G., Delowrth, T., Doblas-Reyes, F. J., Dunstone, N. J., Estella-Perez, V., Flavoni, S., Hermanson, L., Keenlyside, N.,
Kharin, V., Kimoto, M., Merryfield, W. J., Mignot, J., Mochizuki, T., Modali, K., Moneri, P.-A., Müller, W. A., Nicolı́, D., Ortega, P.,
Pankatz, K., Pholman, H., Robson, J., Ruggieri, P., Sospedra-Alfonso, R., Swingedouw, D., Wang, Y., Wild, S., Yeager, S., Yang, X.,
Zhang, L., 2020. North Atlantic climate far more predictable than models imply. Nature 583 (7818), 796–800.
Smith, D. R., 2002. Markov-Switching and stochastic volatility diffusion models of Short-Term interest rates. Journal of Business &
Economic Statistics 20 (2), 183–197.
Smith, J., Wallis, K. F., 2009a. A simple explanation of the forecast combination puzzle. Oxford Bulletin of Economics and Statistics
71 (3), 331–355.
Smith, J., Wallis, K. F., 2009b. A simple explanation of the forecast combination puzzle. Oxford Bulletin of Economics and Etatistics
71 (3), 331–355.
Smith, J. C., 2011. The ins and outs of UK unemployment. The Economic Journal 121, 402–444.
Smith, M., 2010. Modeling Longitudinal Data Using a Pair-Copula Decomposition of Serial Dependence. Journal of the American
Statistical Association 105 (492), 1467–1479.
Smith, M., Khaled, M., 2012. Estimation of copula models with discrete margins via Bayesian data augmentation. Journal of the
American Statistical Association 107 (497), 290–303.
Smith, M. S., Maneesoonthorn, W., 2018. Inversion copulas from nonlinear state space models with an application to inflation fore-
casting. International Journal of Forecasting 34 (3), 389–407.
Smith, M. S., Vahey, S. P., 2016. Asymmetric forecast densities for US macroeconomic variables from a gaussian copula model of
cross-sectional and serial dependence. Journal of Business & Economic Statistics 34 (3), 416–434.
Smith, P., 2016. Google’s MIDAS touch: Predicting UK unemployment with internet search data. Journal of Forecasting 35 (3), 263–
284.
Smyl, S., 2020. A hybrid method of exponential smoothing and recurrent neural networks for time series forecasting. International
Journal of Forecasting 36 (1), 75–85.
Sniezek, J. A., Henry, R. A., 1989. Accuracy and confidence in group judgment. Organizational Behavior and Human Decision Pro-
cesses 43 (1), 1–28.
Sobhani, M., Hong, T., Martin, C., 2020. Temperature anomaly detection for electric load forecasting. International Journal of Fore-
casting 36 (2), 324–333.
Sobri, S., Koohi-Kamali, S., Rahim, N. A., 2018. Solar photovoltaic generation forecasting methods: A review. Energy Conversion &
Management 156, 459–497.
Soebiyanto, R. P., Adimi, F., Kiang, R. K., 2010. Modeling and predicting seasonal influenza transmission in warm regions using

230
 climatological parameters. PloS One 5 (3), e9450.
Sommer, B., Pinson, P., Messner, J. W., Obst, D., 2020. Online distributed learning in wind power forecasting. International Journal
of Forecasting.
Son, N., Yang, S., Na, J., 2019. Hybrid forecasting model for Short-Term wind power prediction using modified long Short-Term
memory. Energies 12 (20), 3901.
Song, H., Qiu, R. T. R., Park, J., 2019. A review of research on tourism demand forecasting: Launching the annals of tourism research
curated collection on tourism demand forecasting. Annals Of Tourism Research 75, 338–362.
Song, H., Witt, S. F., Li, G., 2008. The Advanced Econometrics of Tourism Demand. Routledge.
Sopadjieva, E., Dholakia, U. M., Benjamin, B., 2017. A study of 46,000 shoppers shows that omnichannel retailing works. Harvard
Business Review Reprint H03D7A.
Sorensen, D., 2020. Strategic IBP: Driving profitable growth in complex global organizations. Foresight: The International Journal of
Applied Forecasting 56, 36–45.
Sornette, D., 2003. Critical market crashes. Physics Reports 378 (1), 1–98.
Soule, D., Grushka-Cockayne, Y., Merrick, J. R. W., 2020. A heuristic for combining correlated experts. SSRN:3680229.
Souza, R. C., Marcato, A. L. M., Dias, B. H., Oliveira, F. L. C., 2012. Optimal operation of hydrothermal systems with hydrological
scenario generation through bootstrap and periodic autoregressive models. European Journal of Operational Research 222 (3),
606–615.
Soyer, R., Tarimcilar, M. M., 2008. Modeling and analysis of call center arrival data: A Bayesian approach. Management Science 54 (2),
266–278.
Spagat, M., Mack, A., Cooper, T., Kreutz, J., 2009. Estimating war deaths: An arena of contestation. The Journal of Conflict Resolution
53 (6), 934–950.
Sparkes, J. R., McHugh, A. K., 1984. Awareness and use of forecasting techniques in British industry. Journal of Forecasting 3 (1),
37–42.
Spencer, J., 1904. On the graduation of the rates of sickness and mortality presented by the experience of the Manchester Unity of
Oddfellows during the period 1893-97. Journal of the Institute of Actuaries 38 (4), 334–343.
Spiegelhalter, D., Pearson, M., Short, I., 2011. Visualizing uncertainty about the future. Science 333 (6048), 1393–1400.
Spiliotis, E., Assimakopoulos, V., Makridakis, S., 2020a. Generalizing the theta method for automatic forecasting. European Journal
of Operational Research 284 (2), 550–558.
Spiliotis, E., Assimakopoulos, V., Nikolopoulos, K., 2019a. Forecasting with a hybrid method utilizing data smoothing, a variation of
the theta method and shrinkage of seasonal factors. International Journal of Production Economics 209, 92–102.
Spiliotis, E., Kouloumos, A., Assimakopoulos, V., Makridakis, S., 2020b. Are forecasting competitions data representative of the
reality? International Journal of Forecasting 36 (1), 37–53.
Spiliotis, E., Petropoulos, F., Assimakopoulos, V., 2019b. Improving the forecasting performance of temporal hierarchies. PloS One
14 (10), e0223422.
Spiliotis, E., Petropoulos, F., Kourentzes, N., Assimakopoulos, V., 2020c. Cross-temporal aggregation: Improving the forecast accuracy
of hierarchical electricity consumption. Applied Energy 261, 114339.
Spiliotis, E., Raptis, A., Assimakopoulos, V., 2015. Off-the-shelf vs. customized forecasting support systems. Foresight: The Interna-
tional Journal of Applied Forecasting, issue 43, 42–48.
Spithourakis, G., Petropoulos, F., Babai, M. Z., Nikolopoulos, K., Assimakopoulos, V., 2011. Improving the performance of popular
supply chain forecasting techniques. Supply Chain Forum, an International Journal 12 (4), 16–25.
Spithourakis, G., Petropoulos, F., Nikolopoulos, K., Assimakopoulos, V., 2014. A systemic view of ADIDA framework. IMA Journal of
Management Mathematics 25, 125–137.
Squire, P., 1988. Why the 1936 literary digest poll failed. Public Opinion Quarterly 52 (1), 125–133.
Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., Salakhutdinov, R., 2014. Dropout: A simple way to prevent neural networks
from overfitting. Journal of Machine Learning Research 15 (56), 1929–1958.
Stadlober, E., Hormann, S., Pfeiler, B., 2018. Quality and performance of a PM10 daily forecasting model. Atmospheric Environment
42, 1098–1109.
Stanford NLP Group, 2013. Code for deeply moving: Deep learning for sentiment analysis. https://nlp.stanford.edu/sentiment/
code.html, accessed on 2020-09-05.
Stanton, R., 1997. A nonparametric model of term structure dynamics and the market price of interest rate risk. The Journal of
Finance 52 (5), 1973–2002.
Staszewska-Bystrova, A., 2011. Bootstrap prediction bands for forecast paths from vector autoregressive models. Journal of Forecast-

231
 ing 30 (8), 721–735.
Steckley, S. G., Henderson, S. G., Mehrotra, V., 2005. Performance measures for service systems with a random arrival rate. In:
Proceedings of the Winter Simulation Conference, 2005. IEEE.
Steurer, J., 2011. The delphi method: an efficient procedure to generate knowledge. Skeletal Radiology 40 (8), 959–961.
Stock, J. H., Watson, M. W., 1998. A comparison of linear and nonlinear univariate models for forecasting macroeconomic time series.
NBER Working Papers 6607, National Bureau of Economic Research, Inc.
Stock, J. H., Watson, M. W., 2004. Combination forecasts of output growth in a seven-country data set. Journal of Forecasting 23 (6),
405–430.
Stock, J. H., Watson, M. W., 2012. Generalized shrinkage methods for forecasting using many predictors. Journal of Business &
Economic Statistics 30, 481–493.
Stone, M., 1961. The opinion pool. Annals of Mathematical Statistics 32 (4), 1339–1342.
Stone, M., 1974. Cross-validatory choice and assessment of statistical predictions. Journal of the Royal Statistical Society, Series B
(Statistical Methodology) 36 (2), 111–133.
Storey, M. V., Van der Gaag, B., Burns, B. P., 2011. Advances in on-line drinking water quality monitoring and early warning systems.
Water research 45 (2), 741–747.
Strähl, C., Ziegel, J., 2017. Cross-calibration of probabilistic forecasts. Electronic Journal of Statistics 11 (1), 608–639.
Strauch, R., Hallerberg, M., Hagen, J., 2004. Budgetary forecasts in Europe – the track record of stability and convergence pro-
grammes. ECB Working Paper 307.
Strijbosch, L. W. G., Moors, J. J. A., 2005. The impact of unknown demand parameters on (R,S)-inventory control performance.
European Journal of Operational Research 162 (3), 805–815.
Su, Y.-K., Wu, C.-C., 2014. A new range-based regime-switching dynamic conditional correlation model for minimum-variance hedg-
ing. Journal of Mathematical Finance 04 (03), 207–219.
Sugeno, M., 1985. Industrial applications of fuzzy control. Elsevier Science Inc.
Sun, S., Sun, Y., Wang, S., Wei, Y., 2018. Interval decomposition ensemble approach for crude oil price forecasting. Energy Economics
76, 274–287.
Sundquist, E. T., Keeling, R. F., 2009. The Mauna Loa carbon dioxide record: Lessons for long-term earth observations. Geophysical
Monograph Series 183, 27–35.
Surowiecki, J., 2005. The Wisdom of Crowds: Why the Many Are Smarter Than the Few, new Edition. Abacus.
Svensson, A., Holst, J., Lindquist, R., Lindgren, G., 1996. Optimal prediction of catastrophes in autoregressive Moving-Average
processes. Journal of Time Series Analysis 17 (5), 511–531.
Svetunkov, I., Boylan, J. E., 2020. State-space ARIMA for supply-chain forecasting. International Journal of Production Research
58 (3), 818–827.
Swanson, N. R., Xiong, W., 2018. Big data analytics in economics: What have we learned so far, and where should we go from here?
Canadian Journal of Economics 51 (3), 695–746.
Sweeney, C., Bessa, R. J., Browell, J., Pinson, P., 2019. The future of forecasting for renewable energy. Wiley Interdisciplinary Reviews:
Energy and Environment.
Syntetos, A. A., Babai, M. Z., Luo, S., 2015a. Forecasting of compound Erlang demand. Journal of the Operational Research Society
66 (12), 2061–2074.
Syntetos, A. A., Babai, Z., Boylan, J. E., Kolassa, S., Nikolopoulos, K., 2016a. Supply chain forecasting: Theory, practice, their gap and
the future. European Journal of Operational Research 252 (1), 1–26.
Syntetos, A. A., Boylan, J. E., 2001. On the bias of intermittent demand estimates. International Journal of Production Economics
71 (1), 457–466.
Syntetos, A. A., Boylan, J. E., 2005. The accuracy of intermittent demand estimates. International Journal of Forecasting 21 (2),
303–314.
Syntetos, A. A., Boylan, J. E., 2006. On the stock control performance of intermittent demand estimators. International Journal of
Production Economics 103 (1), 36–47.
Syntetos, A. A., Boylan, J. E., Croston, J. D., 2005. On the categorization of demand patterns. Journal of the Operational Research
Society 56 (5), 495–503.
Syntetos, A. A., Kholidasari, I., Naim, M. M., 2016b. The effects of integrating management judgement into out levels: In or out of
context? European Journal of Operational Research 249 (3), 853–863.
Syntetos, A. A., Nikolopoulos, K., Boylan, J. E., Fildes, R., Goodwin, P., 2009. The effects of integrating management judgement into
intermittent demand forecasts. International Journal of Production Economics 118 (1), 72–81.

232
Syntetos, A. A., Zied Babai, M., Gardner, E. S., 2015b. Forecasting intermittent inventory demands: simple parametric methods vs.
bootstrapping. Journal of Business Research 68 (8), 1746–1752.
Szozda, N., 2010. Analogous forecasting of products with a short life cycle. Decision Making in Manufacturing and Services 4 (1-2),
71–85.
Taillardat, M., Mestre, O., Zamo, M., Naveau, P., 2016. Calibrated ensemble forecasts using quantile regression forests and ensemble
model output statistics. Monthly Weather Review 144 (6), 2375–2393.
Talagala, P. D., Hyndman, R. J., Leigh, C., Mengersen, K., Smith-Miles, K., 2019. A feature-based procedure for detecting technical
outliers in water-quality data from in situ sensors. Water Resources Research 55 (11), 8547–8568.
Talagala, P. D., Hyndman, R. J., Smith-Miles, K., 2020a. Anomaly detection in high dimensional data. Journal of Computational and
Graphical Statistics in press, 1–32.
Talagala, P. D., Hyndman, R. J., Smith-Miles, K., Kandanaarachchi, S., Muñoz, M. A., 2020b. Anomaly detection in streaming nonsta-
tionary temporal data. Journal of Computational and Graphical Statistics 29 (1), 13–27.
Talagala, T. S., Hyndman, R. J., Athanasopoulos, G., 2018. Meta-learning how to forecast time series. Working paper 6/18, Monash
University, Department of Econometrics and Business Statistics.
Talavera-Llames, R. L., Pérez-Chacón, R., Martı́nez-Ballesteros, M., Troncoso, A., Martı́nez-Álvarez, F., 2016. A nearest neighbours-
based algorithm for big time series data forecasting. In: International Conference on Hybrid Artificial Intelligence Systems.
Springer, pp. 174–185.
Taleb, N. N., 2008. The Black Swan: The Impact of the Highly Improbable, new edition Edition. Penguin.
Taleb, N. N., 2020. Statistical Consequences of Fat Tails: Real World Preasymptotics, Epistemology, and Applications. STEM Academic
Press.
Taleb, N. N., Bar-Yam, Y., Cirillo, P., 2020. On single point forecasts for fat tailed variables. International Journal of Forecasting.
Tam Cho, W. K., 1998. Iff the assumption fits...: A comment on the king ecological inference solution. Political Analysis 7, 143–163.
Tan, B. K., Panagiotelis, A., Athanasopoulos, G., 2019. Bayesian inference for the one-factor copula model. Journal of Computational
and Graphical Statistics 28 (1), 155–173.
Tan, P.-N., Steinbach, M., Kumar, V., 2005. Introduction to Data Mining, (First Edition). Addison-Wesley Longman Publishing Co.,
Inc., Boston, MA, USA.
Tandberg, D., Easom, L. J., Qualls, C., 1995. Time series forecasts of poison center call volume. Journal of Toxicology: Clinical
Toxicology 33 (1), 11–18.
Tanir, O., Booth, R. J., 1999. Call center simulation in Bell Canada. In: WSC’99. 1999 Winter Simulation Conference
Proceedings.’Simulation-A Bridge to the Future’(Cat. No. 99CH37038). Vol. 2. IEEE, pp. 1640–1647.
Tarun, G., Bryan, K., 2019. Factor momentum everywhere. Journal of Portfolio Management 45 (3), 13–36.
Tashman, L. J., 2000. Out-of-sample tests of forecasting accuracy: an analysis and review. International Journal of Forecasting 16 (4),
437–450.
Tashman, L. J., Leach, M. L., 1991. Automatic forecasting software: A survey and evaluation. International Journal of Forecasting
7 (2), 209–230.
Tay, A. S., Wallis, K. F., 2000. Density forecasting: a survey. Journal of Forecasting 19 (4), 235–254.
Taylor, J., Jeon, J., 2018. Probabilistic forecasting of wave height for offshore wind turbine maintenance. European Journal of Opera-
tional Research 267 (3).
Taylor, J. W., 2003a. Exponential smoothing with a damped multiplicative trend. International Journal of Forecasting 19 (4), 715–725.
Taylor, J. W., 2003b. Short-term electricity demand forecasting using double seasonal exponential smoothing. Journal of the Opera-
tional Research Society 54 (8), 799–805.
Taylor, J. W., 2007. Forecasting daily supermarket sales using exponentially weighted quantile regression. European Journal of Oper-
ational Research 178 (1), 154–167.
Taylor, J. W., 2008. A comparison of univariate time series methods for forecasting intraday arrivals at a call center. Management
Science 54 (2), 253–265.
Taylor, J. W., 2010. Exponentially weighted methods for forecasting intraday time series with multiple seasonal cycles. International
Journal of Forecasting 26 (4), 627–646.
Taylor, J. W., 2012. Density forecasting of intraday call center arrivals using models based on exponential smoothing. Management
Science 58 (3), 534–549.
Taylor, J. W., Bunn, D. W., 1999. A quantile regression approach to generating prediction intervals. Management Science 45 (2),
131–295.
Taylor, J. W., McSharry, P. E., Buizza, R., 2009. Wind Power Density Forecasting Using Ensemble Predictions and Time Series Models.

233
 IEEE Transactions on Energy Conversion 24 (3), 775–782.
Taylor, J. W., Snyder, R. D., 2012. Forecasting intraday time series with multiple seasonal cycles using parsimonious seasonal expo-
nential smoothing. Omega 40 (6), 748–757.
Taylor, M. P., Peel, D. A., 2000. Nonlinear adjustment, long-run equilibrium and exchange rate fundamentals. Journal of International
Money and Finance 19 (1), 33–53.
Taylor, P. F., Thomas, M. E., 1982. Short term forecasting: Horses for courses. Journal of the Operational Research Society 33 (8),
685–694.
Taylor, S., 1986. Modelling Financial Time Series. Wiley.
Taylor, S. J., Letham, B., 2018. Forecasting at scale. The American Statistician 72 (1), 37–45.
Tenti, P., 1996. Forecasting foreign exchange rates using recurrent neural networks. Applied Artificial Intelligence 10 (6), 567–582.
Teräsvirta, T., 1994. Specification, estimation, and evaluation of smooth transition autoregressive models. Journal of the American
Statistical Association 89 (425), 208–218.
Tetlock, P., Gardner, D., 2015. Superforecasting: The Art and Science of Prediction. Crown, New York.
Teunter, R. H., Duncan, L., 2009. Forecasting intermittent demand: a comparative study. Journal of the Operational Research Society
60 (3), 321–329.
Teunter, R. H., Syntetos, A. A., Zied Babai, M., 2011. Intermittent demand: Linking forecasting to inventory obsolescence. European
Journal of Operational Research 214 (3), 606–615.
Tewari, D. D., 1990. Energy-price impacts modelling in the agriculture sector. Energy Economics 12 (2), 147–158.
Thaler, R. H., Sunstein, C. R., 2009. Nudge: Improving Decisions about Health, Wealth, and Happiness. Penguin Books.
The Conference Board, 2020. Global business cycle indicators. https://conference-board.org/data/bcicountry.cfm?cid=1, ac-
cessed on 2020-09-07.
The RFE Working Group Report, 2015. Risk in the front end of megaprojects. Tech. rep., European Cooperation in Science and
Technology.
Theocharis, Z., Harvey, N., 2019. When does more mean worse? Accuracy of judgmental forecasting is nonlinearly related to length
of data series. Omega 87, 10–19.
Theocharis, Z., Smith, L. A., Harvey, N., 2018. The influence of graphical format on judgmental forecasting accuracy: Lines versus
points. Futures & Foresight Science 13, e7.
Theodosiou, M., 2011. Disaggregation & aggregation of time series components: A hybrid forecasting approach using generalized
regression neural networks and the theta method. Neurocomputing 74 (6), 896–905.
Thomakos, D., Nikolopoulos, K., 2012. Fathoming the theta method for a unit root process. IMA Journal of Management Mathematics
25 (1), 105–124.
Thomakos, D. D., Nikolopoulos, K., 2015. Forecasting multivariate time series with the theta method: Multivariate theta method.
Journal of Forecasting 34 (3), 220–229.
Thomé, A. M. T., Hollmann, R. L., Scavarda do Carmo, L. F. R. R., 2014. Research synthesis in collaborative planning forecast and
replenishment. Industrial Management & Data Systems 114 (6), 949–965.
Thomé, A. M. T., Scavarda, L. F., Fernandez, N. S., Scavarda, A. J., 2012. Sales and operations planning: A research synthesis. Inter-
national Journal of Production Economics 138 (1), 1–13.
Thomson, W., Jabbari, S., Taylor, A., Arlt, W., Smith, D., 2019. Simultaneous parameter estimation and variable selection via the
logit-normal continuous analogue of the spike-and-slab prior. Journal of the Royal Society Interface 16 (150).
Thorarinsdottir, T. L., Scheuerer, M., Heinz, C., 2016. Assessing the calibration of high-dimensional ensemble forecasts using rank
histograms. Journal of Computational and Graphical Statistics 25 (1), 105–122.
Thorarinsdottir, T. L., Schuhen, N., 2018. Verification: assessment of calibration and accuracy. In: Statistical Postprocessing of En-
semble Forecasts. Elsevier, pp. 155–186.
Tian, F., Yang, K., Chen, L., 2017. Realized volatility forecasting of agricultural commodity futures using the HAR model with time-
varying sparsity. International Journal of Forecasting 33 (1), 132–152.
Tian, J., Anderson, H. M., 2014. Forecast combinations under structural break uncertainty. International Journal of Forecasting 30 (1),
161–175.
Tibshirani, R., 1996. Regression shrinkage and selection via the LASSO. Journal of the Royal Statistical Society, Series B (Statistical
Methodology) 58, 267–288.
Timmermann, A., 2000. Moments of Markov switching models. Journal of Econometrics 96 (1), 75–111.
Timmermann, A., 2006. Forecast combinations. In: Elliott, G., Granger, C. W. J., Timmermann, A. (Eds.), Handbook of Economic
Forecasting. Vol. 1. Elsevier, Amsterdam, pp. 135–196.

234
Timmermann, A., Zhu, Y., 2019. Comparing forecasting performance with panel data. SSRN:3380755.
Tiwari, A. K., Nasreen, S., Shahbaz, M., Hammoudeh, S., 2020. Time-frequency causality and connectedness between international
prices of energy, food, industry, agriculture and metals. Energy Economics 85, 104529.
Todini, E., 1991. Coupling real-time forecasting in the aswan dam reservoir management. In: Workshop on Monitoring, Forecasting
and Simulation of River Basins for Agricultural Production. Rome.
Todini, E., 1999. Using phase-space modelling for inferring forecasting uncertainty in non-linear stochastic decision schemes. Journal
of Hydroinformatics 1 (2), 75–82.
Todini, E., 2008. A model conditional processor to assess predictive uncertainty in flood forecasting. International Journal of River
Basin Management 6 (2), 123–137.
Todini, E., 2016. Predictive uncertainty assessment and decision making. In: Singh, V. P. (Ed.), Handbook of Applied Hydrology.
McGraw Hill, New York, pp. 26.1–26.16.
Todini, E., 2017. Flood forecasting and decision making in the new millennium. Where are we? Water Resources Management 31 (10),
3111–3129.
Todini, E., 2018. Paradigmatic changes required in water resources management to benefit from probabilistic forecasts. Water Security
3, 9–17.
Toktay, L. B., 2003. Forecasting product returns. In: Guide, Jr., V. D. R., van Wassenhove, L. N. (Eds.), Business aspects of closed-loop
supply chains. Carnegie Mellon University Press, Pittsburgh, pp. 203–209.
Toktay, L. B., Wein, L. M., Zenios, S. A., 2000. Inventory management of remanufacturable products. Management Science 46 (11),
1412–1426.
Tolman, H. L., 2008. A mosaic approach to wind wave modeling. Ocean Modelling 25 (1-2), 35–47.
Tong, H., 1978. On a threshold model. In: Chen, C. (Ed.), Pattern Recognition and Signal Processing. NATO ASI Series E: Applied Sc.
Sijthoff & Noordhoff, Netherlands, pp. 575–586.
Toth, Z., Buizza, R., 2019. Weather forecasting: What sets the forecast skill horizon? In: Sub-Seasonal to Seasonal Prediction. Elsevier,
pp. 17–45.
Touzani, S., Granderson, J., Fernandes, S., 2018. Gradient boosting machine for modeling the energy consumption of commercial
buildings. Energy and Buildings 158, 1533–1543.
Tracy, M., Cerdá, M., Keyes, K. M., 2018. Agent-based modeling in public health: current applications and future directions. Annual
Review of Public Health 39, 77–94.
Tran, T., Phung, D., Luo, W., Harvey, R., Berk, M., Venkatesh, S., 2013. An integrated framework for suicide risk prediction. In:
Proceedings of the 19th ACM SIGKDD international conference on Knowledge Discovery and Data Mining. pp. 1410–1418.
Trapero, J. R., Kourentzes, N., Fildes, R., 2015. On the identification of sales forecasting models in the presence of promotions. Journal
of the operational Research Society 66 (2), 299–307.
Trapero, J. R., Pedregal, D. J., Fildes, R., Kourentzes, N., 2013. Analysis of judgmental adjustments in the presence of promotions.
International Journal of Forecasting 29 (2), 234–243.
Triguero, I., Peralta, D., Bacardit, J., Garcı́a, S., Herrera, F., 2015. MRPR: A MapReduce solution for prototype reduction in big data
classification. Neurocomputing 150, 331–345.
Trivedi, P. K., Zimmer, D. M., 2007. Copula modeling: an introduction for practitioners. Now Publishers Inc.
Tsay, R. S., 1986. Time series model specification in the presence of outliers. Journal of the American Statistical Association 81 (393),
132–141.
Tse, Y. K., Tsui, A. K. C., 2002. A Multivariate Generalized Autoregressive Conditional Heteroscedasticity Model With Time-Varying
Correlations. Journal of Business & Economic Statistics 20 (3), 351–362.
Tsyplakov, A., 2013. Evaluation of probabilistic forecasts: proper scoring rules and moments. SSRN:2236605.
Tu, Y., Ball, M., Jank, W., 2008. Estimating flight departure delay distributions – a statistical approach with long-term trend and
short-term pattern. Journal of the American Statistical Association 103 (481), 112–125.
Turkman, M. A. A., Turkman, K. F., 1990. Optimal alarm systems for autoregressive processes: A Bayesian approach. Computational
Statistics & Data Analysis 10 (3), 307–314.
Turkmen, A. C., Wang, Y., Januschowski, T., 2019. Intermittent demand forecasting with deep renewal processes. arXiv:1911.10416.
Turner, D. S., 1990. The role of judgement in macroeconomic forecasting. Journal of Forecasting 9 (4), 315–345.
Turner, L., Boulhol, H., 2011. Recent trends and structural breaks in the US and EU15 labour productivity growth. Applied Economics
43 (30), 4769–4784.
Turner, R., Zolin, R., 2012. Forecasting success on large projects: Developing reliable scales to predict multiple perspectives by
multiple stakeholders over multiple time frames. Project Management Journal 43 (5), 87–99.

235
Turnovsky, S. J., Wachter, M. L., 1972. A Test of the “Expectations Hypothesis” Using Directly Observed Wage and Price Expectations.
The Review of Economics and Statistics 54 (1), 47–54.
Twyman, M., Harvey, N., Harries, C., 2008. Trust in motives, trust in competence: Separate factors determining the effectiveness of
risk communication. Judgment and Decision Making 3 (1), 111–120.
Tych, W., Pedregal, D. J., Young, P. C., Davies, J., 2002. An unobserved component model for multi-rate forecasting of telephone call
demand: the design of a forecasting support system. International Journal of forecasting 18 (4), 673–695.
Tziafetas, G., 1986. Estimation of the voter transition matrix. Optimization 17 (2), 275–279.
Uematsu, H., Kunisawa, S., Sasaki, N., Ikai, H., Imanaka, Y., 2014. Development of a risk-adjusted in-hospital mortality prediction
model for community-acquired pneumonia: a retrospective analysis using a Japanese administrative database. BMC Pulmonary
Medicine 14 (1), 203.
Ugurlu, U., Oksuz, I., Tas, O., 2018. Electricity price forecasting using recurrent neural networks. Energies 11 (5), 1255.
Ülkümen, G., Fox, C. R., Malle, B. F., 2016. Two dimensions of subjective uncertainty: Clues from natural language. Journal of
Experimental Psychology: General 145 (10), 1280–1297.
Unwin, A., 2019. Multivariate outliers and the O3 Plot. Journal of Computational and Graphical Statistics 28 (3), 635–643.
Vaks, A., Mason, A. J., Breitenbach, S. F. M., et al., 2019. Palaeoclimate evidence of vulnerable permafrost during times of low sea ice.
Nature 577, 221–225.
Van de Ven, A., Delbeco, A. L., 1971. Nominal versus interacting group processes for committee Decision-Making effectiveness.
Academy of Management journal. Academy of Management 14 (2), 203–212.
Van den Broeke, M., De Baets, S., Vereecke, A., Baecke, P., Vanderheyden, K., 2019. Judgmental forecast adjustments over different
time horizons. Omega 87, 34–45.
van den Hengel, G., Franses, P. H., 2020. Forecasting social conflicts in Africa using an epidemic type aftershock sequence model.
Forecasting 2 (3), 284–308.
Van der Auweraer, S., Boute, R., 2019. Forecasting spare part demand using service maintenance information. International Journal
of Production Economics 213, 138–149.
van der Bles, A. M., van der Linden, S., Freeman, A. L., Mitchell, J., Galvao, A. B., Zaval, L., Spiegelhalter, D. J., 2019. Communicating
uncertainty about facts, numbers and science. Royal Society Open Science 6 (5), 181870.
van der Bles, A. M., van der Linden, S., Freeman, A. L., Spiegelhalter, D. J., 2020. The effects of communicating uncertainty on public
trust in facts and numbers. Proceedings of the National Academy of Sciences 117 (14), 7672–7683.
van der Laan, E., van Dalen, J., Rohrmoser, M., Simpson, R., 2016. Demand forecasting and order planning for humanitarian logistics:
An empirical assessment. Journal of Operations Management 45, 114–122.
van der Mark, L. B., van Wonderen, K. E., Mohrs, J., van Aalderen, W. M., ter Riet, G., Bindels, P. J., 2014. Predicting asthma in
preschool children at high risk presenting in primary care: development of a clinical asthma prediction score. Primary Care
Respiratory Journal 23 (1), 52–59.
Van Dijk, D., Franses, P. H., Lucas, A., 1999. Testing for smooth transition nonlinearity in the presence of outliers. Journal of Business
& Economic Statistics 17 (2), 217–235.
van Donselaar, K., Peters, J., de Jong, A., Broekmeulen, R., 2016. Analysis and forecasting of demand during promotions for perishable
items. International Journal of Production Economics 172, 65–75.
Van Heerde, H. J., Leeflang, P. S., Wittink, D. R., 2002. How promotions work: SCAN* PRO-based evolutionary model building.
Schmalenbach Business Review 54 (3), 198–220.
Van Schaeybroeck, B., Vannitsem, S., 2018. Postprocessing of long-range forecasts. In: Vannitsem, S., Wilks, D. S., Messner, J. W.
(Eds.), Statistical Postprocessing of Ensemble Forecasts. Elsevier, Amsterdam, Netherlands, Ch. 10, pp. 267–290.
van Wingerden, E., Basten, R. J. I., Dekker, R., Rustenburg, W. D., 2014. More grip on inventory control through improved forecasting:
A comparative study at three companies. International Journal of Production Economics 157, 220–237.
Vannitsem, S., Wilks, D. S., Messner, J., 2018. Statistical postprocessing of ensemble forecasts. Elsevier.
Vargas, S. A., Esteves, G. R. T., Maçaira, P. M., Bastos, B. Q., Cyrino Oliveira, F. L., Souza, R. C., 2019. Wind power generation: A
review and a research agenda. Journal of Cleaner Production 218, 850–870.
Varian, H. R., 2014. Big data: New tricks for econometrics. Journal of Economic Perspectives 28 (2), 3–28.
Vaughan Williams, L., Reade, J., 2016. Prediction Markets, Social Media and Information Efficiency. Kyklos 69 (3), 518–556.
Venkatramanan, S., Lewis, B., Chen, J., Higdon, D., Vullikanti, A., Marathe, M., 2018. Using data-driven agent-based models for
forecasting emerging infectious diseases. Epidemics 22, 43–49.
Venter, J. H., De Jongh, P. J., Griebenow, G., 2005. NIG-Garch models based on open, close, high and low prices. South African
Statistical Journal 39 (2), 79–101.

236
Verhulst, P., 1838. Notice sur la loi que la population suit dans son accroissement. Correspondance Mathématique et Physique 10,
113–121.
Vermue, M., Seger, C. R., Sanfey, A. G., 2018. Group-based biases influence learning about individual trustworthiness. Journal of
Experimental Social Psychology 77, 36–49.
Veronesi, P., Yared, F., 1999. Short and long horizon term and inflation risk premia in the US term structure: Evidence from an
integrated model for nominal and real bond prices under regime shifts. CRSP Working Paper 508.
Vestergaard, L. S., Nielsen, J., Richter, L., Schmid, D., Bustos, N., Braeye, T., Denissov, G., Veideman, T., Luomala, O., Möttönen,
T., Fouillet, A., Caserio-Schönemann, C., an der Heiden, M., Uphoff, H., Lytras, T., Gkolfinopoulou, K., Paldy, A., Domegan, L.,
O’Donnell, J., de’ Donato, F., Noccioli, F., Hoffmann, P., Velez, T., England, K., van Asten, L., White, R. A., Tønnessen, R., da Silva,
S. P., Rodrigues, A. P., Larrauri, A., Delgado-Sanz, C., Farah, A., Galanis, I., Junker, C., Perisa, D., Sinnathamby, M., Andrews, N.,
O’Doherty, M., Marquess, D. F., Kennedy, S., Olsen, S. J., Pebody, R., ECDC Public Health Emergency Team for COVID-19, Krause,
T. G., Mølbak, K., 2020. Excess all-cause mortality during the COVID-19 pandemic in Europe–preliminary pooled estimates from
the EuroMOMO network, March to April 2020. Eurosurveillance 25 (26), 2001214.
Vile, J. L., Gillard, J. W., Harper, P. R., Knight, V. A., 2012. Predicting ambulance demand using singular spectrum analysis. Journal
of the Operational Research Society 63 (11), 1556–1565.
Villegas, M. A., Pedregal, D. J., 2018. Supply chain decision support systems based on a novel hierarchical forecasting approach.
Decision Support Systems 114, 29–36.
Vipul, Jacob, J., 2007. Forecasting performance of extreme-value volatility estimators. Journal of Futures Markets 27 (11), 1085–1105.
Vitart, F., Robertson, A. W., Anderson, D. L. T., 2012. Subseasonal to seasonal prediction project: Bridging the gap between weather
and climate. Bulletin of the World Meteorological Organization 61 (2), 23.
Vlahogianni, E. I., Golias, J. C., Karlaftis, M. G., 2004. Short-term traffic forecasting: Overview of objectives and methods. Transport
Reviews 24 (5), 533–557.
Vlahogianni, E. I., Karlaftis, M. G., Golias, J. C., 2014. Short-term traffic forecasting: Where we are and where we’re going. Trans-
portation Research Part C: Emerging Technologies 43, 3–19.
Volterra, V., 1926. Fluctuations in the abundance of a species considered mathematically. Nature 118, 558–60.
Wagner, M. M., Tsui, F.-C., Espino, J. U., Dato, V. M., Sittig, D. F., Caruana, R. A., McGinnis, L. F., Deerfield, D. W., Druzdzel, M. J.,
Fridsma, D. B., 2001. The emerging science of very early detection of disease outbreaks. Journal of Public Health Management and
Practice 7 (6), 51–59.
Wakefield, J., 2004. Ecological inference for 2x2 tables (with discussion). Journal of the Royal Statistical Society, Series A 167 (3),
385–445.
Wallentin, G., Kaziyeva, D., Reibersdorfer-Adelsberger, E., 2020. COVID-19 intervention scenarios for a long-term disease manage-
ment. International Journal of Health Policy and Management.
Wallström, P., Segerstedt, A., 2010. Evaluation of forecasting error measurements and techniques for intermittent demand. Interna-
tional Journal of Production Economics 128 (2), 625–636.
Walton, D., Reed, C., Macagno, F., 2008. Argumentation Schemes. Cambridge University Press.
Wang, C., Jiang, B., Fan, J., Wang, F., Liu, Q., 2014. A study of the dengue epidemic and meteorological factors in Guangzhou, China,
by using a zero-inflated poisson regression model. Asia Pacific Journal of Public Health 26 (1), 48–57.
Wang, H., Li, B., Leng, C., 2009a. Shrinkage tuning parameter selection with a diverging number of parameters. Journal of the Royal
Statistical Society, Series B (Statistical Methodology) 71 (3), 671–683.
Wang, H., Yeung, D.-Y., 2016. A survey on Bayesian deep learning. arXiv:1604.01662.
Wang, J., Wang, Z., Li, X., Zhou, H., 2019a. Artificial bee colony-based combination approach to forecasting agricultural commodity
prices. International Journal of Forecasting.
Wang, J., Yang, W., Du, P., Niu, T., 2018a. A novel hybrid forecasting system of wind speed based on a newly developed multi-objective
sine cosine algorithm. Energy Conversion and Management 163, 134–150.
Wang, J., Yang, W., Du, P., Niu, T., 2020a. Outlier-robust hybrid electricity price forecasting model for electricity market management.
Journal of Cleaner Production 249, 119318.
Wang, P., Liu, B., Hong, T., 2016. Electric load forecasting with recency effect: A big data approach. International Journal of Forecast-
ing 32 (3), 585–597.
Wang, S. L., McPhail, L., 2014. Impacts of energy shocks on US agricultural productivity growth and commodity prices—a structural
VAR analysis. Energy Economics 46 (C), 435–444.
Wang, W., Pedrycz, W., Liu, X., 2015a. Time series long-term forecasting model based on information granules and fuzzy clustering.
Engineering Applications of Artificial Intelligence 41, 17–24.

237
Wang, W., Rothschild, D., Goel, S., Gelman, A., 2015b. Forecasting elections with non-representative polls. International Journal of
Forecasting 31 (3), 980–991.
Wang, W., Syntetos, A. A., 2011. Spare parts demand: Linking forecasting to equipment maintenance. Transportation Research Part
E: Logistics and Transportation Review 47 (6), 1194–1209.
Wang, X., Kang, Y., Hyndman, R. J., Li, F., 2020b. Distributed ARIMA models for ultra-long time series. arXiv:2007.09577.
Wang, X., Kang, Y., Petropoulos, F., Li, F., 2020c. Que será será? The uncertainty estimation of feature-based time series forecasts.
arXiv:1908.02891.
Wang, X., Petropoulos, F., 2016. To select or to combine? The inventory performance of model and expert forecasts. International
Journal of Production Research 54 (17), 5271–5282.
Wang, X., Smith-Miles, K., Hyndman, R. J., 2006. Characteristic-based clustering for time series data. Data Mining and Knowledge
Discovery 13 (3), 335–364.
Wang, X., Smith-Miles, K., Hyndman, R. J., 2009b. Rule induction for forecasting method selection: meta-learning the characteristics
of univariate time series. Neurocomputing 72 (10-12), 2581–2594.
Wang, Y., Smola, A., Maddix, D., Gasthaus, J., Foster, D., Januschowski, T., 2019b. Deep factors for forecasting. In: International
Conference on Machine Learning. pp. 6607–6617.
Wang, Z., Wang, W., Liu, C., Wang, Z., Hou, Y., 2017. Probabilistic forecast for multiple wind farms based on regular vine copulas.
IEEE Transactions on Power Systems 33 (1), 578–589.
Wang, Z., Wang, Y., Zeng, R., Srinivasan, R. S., Ahrentzen, S., 2018b. Random forest based hourly building energy prediction. Energy
and Buildings 171, 11–25.
Warne, A., Coenen, G., Christoffel, K., 2010. Forecasting with DSGE models. Working Paper Series 1185, European Central Bank.
Warrender, C., Forrest, S., Pearlmutter, B., 1999. Detecting intrusions using system calls: Alternative data models. In: Proceedings of
the 1999 IEEE Symposium on Security and Privacy (Cat. No. 99CB36344). IEEE, pp. 133–145.
Weaver, W. T., 1971. The Delphi forecasting method. The Phi Delta Kappan 52 (5), 267–271.
Webby, R., O’Connor, M., Edmundson, B., 2005. Forecasting support systems for the incorporation of event information: An empirical
investigation. International Journal of Forecasting 21 (3), 411–423.
Wei, N., Li, C., Peng, X., Zeng, F., Lu, X., 2019. Conventional models and artificial intelligence-based models for energy consumption
forecasting: A review. Journal of Petroleum Science and Engineering 181, 106187.
Wei, W., Hansen, M., 2006. An aggregate demand model for air passenger traffic in the hub-and-spoke network. Transportation
Research Part A: Policy and Practice 40 (10), 841–851.
Wei Su, C., Wang, X.-Q., Tao, R., Oana-Ramona, L., 2019. Do oil prices drive agricultural commodity prices? Further evidence in a
global bio-energy context. Energy 172, 691–701.
Weinberg, J., Brown, L. D., Stroud, J. R., 2007. Bayesian forecasting of an inhomogeneous poisson process with applications to call
center data. Journal of the American Statistical Association 102 (480), 1185–1198.
Weiß, G. N., Supper, H., 2013. Forecasting liquidity-adjusted intraday value-at-risk with vine copulas. Journal of Banking & Finance
37 (9), 3334–3350.
Wen, R., Torkkola, K., Narayanaswamy, B., Madeka, D., 2017. A multi-horizon quantile recurrent forecaster. arXiv:1711.11053.
Weron, R., 2014. Electricity price forecasting: A review of the state-of-the-art with a look into the future. International Journal of
Forecasting 30 (4), 1030–1081.
West, K. D., 1996. Asymptotic inference about predictive ability. Econometrica, 1067–1084.
White, H., 2000. A reality check for data snooping. Econometrica 68 (5), 1097–1126.
Whitt, W., Zhang, X., 2019. Forecasting arrivals and occupancy levels in an emergency department. Operations Research for Health
Care 21, 1–18.
Whittaker, J., Garside, S., Lindveld, K., 1997. Tracking and predicting a network traffic process. International Journal of Forecasting
13 (1), 51–61.
Wicke, L., Dhami, M. K., Önkal, D., Belton, I. K., 2019. Using scenarios to forecast outcomes of a refugee crisis. International Journal
of Forecasting.
Wickramasuriya, S. L., Athanasopoulos, G., Hyndman, R. J., 2019. Optimal forecast reconciliation for hierarchical and grouped time
series through trace minimization. Journal of the American Statistical Association 114 (526), 804–19.
Wilkd, D. S., 2005. Statistical methods in the atmospheric sciences, 2nd Edition. Science & Technology. Elsevier.
Wilks, D. S., 2004. The minimum spanning tree histogram as verification tool for multidimensional ensemble forecasts. Montly
Weather Review 132, 1329–1340.
Wilks, D. S., 2019. Indices of rank histogram flatness and their sampling properties. Monthly Weather Review 147 (2), 763–769.

238
Willemain, T. R., Smart, C. N., Schwarz, H. F., 2004. A new approach to forecasting intermittent demand for service parts inventories.
International Journal of Forecasting 20 (3), 375–387.
Williams, L. V., Reade, J. J., 2016. Forecasting elections. Journal of Forecasting 35 (4).
Winkler, R. L., 1972. A decision-theoretic approach to interval estimation. Journal of the American Statistical Association 67 (337),
187–191.
Winkler, R. L., Grushka-Cockayne, Y., Lichtendahl, Jr, K. C., Jose, V. R. R., 2019. Probability forecasts and their combination: A
research perspective. Decision Analysis 16 (4), 239–260.
Winkler, R. L., Muñoz, J., Cervera, J. L., Bernardo, J. M., Blattenberger, G., Kadane, J. B., Lindley, D. V., Murphy, A. H., Oliver, R. M.,
Rı́os-Insua, D., 1996. Scoring rules and the evaluation of probabilities. Test 5 (1), 1–60.
Wold, H., 1966. Estimation of principal components and related models by iterative least squares. In: Krishnajah, P. R. (Ed.), Multi-
variate Analysis. Academic Press, New York, pp. 391–420.
Wolfers, J., Zitzewitz, E., 2004. Prediction markets. The Journal of Economic Perspectives 18 (2), 107–126.
Wolpert, D. H., Macready, W. G., 1997. No free lunch theorems for optimization. IEEE Transactions on Evolutionary Computation
1 (1), 67–82.
Wolters, M. H., 2015. Evaluating Point and Density Forecasts of DSGE Models. Journal of Applied Econometrics 30 (1), 74–96.
Wong, B. K., Bodnovich, T. A., Selvi, Y., 1995. A bibliography of neural network business applications research: 1988-September
1994. Expert Systems 12 (3), 253–261.
Wong-Fupuy, C., Haberman, S., 2004. Projecting mortality trends: recent developments in the United Kingdom and the United States.
North American Actuarial Journal 8 (2), 56–83.
Woodford, M., 2002. Imperfect common knowledge and the effects of monetary policy. In: Aghion, P., Frydman, R., Stiglitz, J., Wood-
ford, M. (Eds.), Knowledge, Information, and Expectations in Modern Macroeconomics: In honor of Edmund Phelps. Princeton
University Press, pp. 25–58.
Wright, G., Goodwin, P., 1999. Future-focussed thinking: combining scenario planning with decision analysis. Journal of Multi-
Criteria Decision Analysis 8 (6), 311–321.
Wright, G., Goodwin, P., 2009. Decision making and planning under low levels of predictability: Enhancing the scenario method.
International Journal of Forecasting 25 (4), 813–825.
Wright, M. J., Stern, P., 2015. Forecasting new product trial with analogous series. Journal of Business Research 68 (8), 1732–1738.
Wu, C. C., Liang, S. S., 2011. The economic value of range-based covariance between stock and bond returns with dynamic copulas.
Journal of Empirical Finance 18 (4), 711–727.
Wu, S., Chen, R., 2007. Threshold variable determination and threshold variable driven switching autoregressive models. Statistica
Sinica 17 (1), 241–S38.
Xiao, Y., Han, J., 2016. Forecasting new product diffusion with agent-based models. Technological Forecasting and Social Change 105,
167–178.
Xie, J., Hong, T., 2016. GEFCom2014 probabilistic electric load forecasting: An integrated solution with forecast combination and
residual simulation. International Journal of Forecasting 32 (3), 1012–1016.
Xie, T., Ding, J., 2020. Forecasting with multiple seasonality. arXiv:2008.12340.
Xie, W., Yu, L., Xu, S., Wang, S., 2006. A new method for crude oil price forecasting based on support vector machines. In: Interna-
tional Conference on Computational Science. Springer, pp. 444–451.
Xie, Y., Hu, P., Zhu, N., Lei, F., Xing, L., Xu, L., Sun, Q., 2020. A hybrid short-term load forecasting model and its application in
ground source heat pump with cooling storage system. Renewable Energy 161, 1244–1259.
Xiong, T., Li, C., Bao, Y., Hu, Z., Zhang, L., 2015. A combination method for interval forecasting of agricultural commodity futures
prices. Knowledge-Based Systems 77, 92–102.
Xu, R., Wunsch, 2nd, D., 2005. Survey of clustering algorithms. IEEE Transactions on Neural Networks 16 (3), 645–678.
Xu, W., 1999. Long range planning for call centers at Fedex. The Journal of Business Forecasting 18 (4), 7.
Xu, Y., Liu, H., Long, Z., 2020. A distributed computing framework for wind speed big data forecasting on Apache Spark. Sustainable
Energy Technologies and Assessments 37, 100582.
Yaari, M. E., 1965. Uncertain lifetime, life insurance, and the theory of the consumer. The Review of Economic Studies 32 (2), 137–150.
Yaffee, R. A., Nikolopoulos, K., Reilly, D. P., Crone, S. F., Wagoner, K. D., Douglas, R. J., Amman, B. R., Ksiazek, T. G., Mills, J. N.,
2011. An experiment in epidemiological forecasting: A comparison of forecast accuracies among different methods of forecasting
deer mouse population densities in montana.
Yagli, G. M., Yang, D., Srinivasan, D., 2019. Reconciling solar forecasts: Sequential reconciliation. Solar Energy 179, 391–397.
Yang, D., Zhang, Q., 2000. Drift-independent volatility estimation based on high, low, open, and close prices. Journal of Business

239
 73 (3), 477–491.
Yang, Z., Zeng, Z., Wang, K., Wong, S.-S., Liang, W., Zanin, M., Liu, P., Cao, X., Gao, Z., Mai, Z., Liang, J., Liu, X., Li, S., Li, Y., Ye,
F., Guan, W., Yang, Y., Li, F., Luo, S., Xie, Y., Liu, B., Wang, Z., Zhang, S., Wang, Y., Zhong, N., He, J., 2020. Modified SEIR and
AI prediction of the epidemics trend of COVID-19 in china under public health interventions. Journal of Thoracic Disease 12 (3),
165–174.
Yassine, A., Shirehjini, A. N., Shirmohammadi, S., 2015. Smart meters big data: Game theoretic model for fair data sharing in
deregulated smart grids. IEEE Access 3, 2743–2754.
Yelland, P., Baz, Z. E., Serafini, D., 2019. Forecasting at scale: The architecture of a modern retail forecasting system. Foresight: The
International Journal of Applied Forecasting 55, 10–18.
Yusupova, A., Pavlidis, E., Paya, I., Peel, D., 2020. UK housing price uncertainty index (HPU). UK Housing Observatory, Dept. of
Economics, Lancaster University Management School.
Yusupova, A., Pavlidis, N. G., Pavlidis, E. G., 2019. Adaptive dynamic model averaging with an application to house price forecasting.
arXiv:1912.04661.
Zadeh, L., 1965. Fuzzy sets. Information and Control 8 (3), 338–353.
Zagdański, A., 2001. Prediction intervals for stationary time series using the sieve bootstrap method. Demonstratio Mathematica
34 (2), 257–270.
Zaharia, M., Xin, R. S., Wendell, P., Das, T., Armbrust, M., Dave, A., Meng, X., Rosen, J., Venkataraman, S., Franklin, M. J., Ghodsi,
A., Gonzalez, J., Shenker, S., Stoica, I., 2016. Apache Spark: A unified engine for big data processing. Communications of the ACM
59 (11), 56–65.
Zaki, M. J., 2000. Scalable algorithms for association mining. IEEE Transactions on Knowledge and Data Engineering 12 (3), 372–390.
Zakoian, J.-M., 1994. Threshold heteroskedastic models. Journal of Economic Dynamics and Control 18 (5), 931–955.
Zang, H., Cheng, L., Ding, T., Cheung, K. W., Liang, Z., Wei, Z., Sun, G., 2018. Hybrid method for short-term photovoltaic power
forecasting based on deep convolutional neural network. IET Generation, Transmission and Distribution 12 (20), 4557–4567.
Zarnowitz, V., 1985. Rational expectations and macroeconomic forecasts. Journal of Business & Economic Statistics 3 (4), 293–311.
Zelterman, D., 1993. A semiparametric bootstrap technique for simulating extreme order statistics. Journal of the American Statistical
Association 88 (422), 477–485.
Zhang, Ming, 2008. Artificial Higher Order Neural Networks for Economics and Business. IGI Global.
Zhang, G., Eddy Patuwo, B., Y. Hu, M., 1998. Forecasting with artificial neural networks:: The state of the art. International Journal
of Forecasting 14 (1), 35–62.
Zhang, G. P., Qi, M., 2005. Neural network forecasting for seasonal and trend time series. European Journal of Operational Research
160 (2), 501–514.
Zhang, J., Wei, Y.-M., Li, D., Tan, Z., Zhou, J., 2018. Short term electricity load forecasting using a hybrid model. Energy 158, 774–781.
Zhang, J.-L., Chen, J., Lee, C.-Y., 2008. Joint optimization on pricing, promotion and inventory control with stochastic demand.
International Journal of Production Economics 116 (2), 190–198.
Zhang, W., Qi, Y., Henrickson, K., Tang, J., Wang, Y., 2017. Vehicle traffic delay prediction in ferry terminal based on bayesian multiple
models combination method. Transportmetrica A: Transport Science 13 (5), 467–490.
Zhang, W., Zou, Y., Tang, J., Ash, J., Wang, Y., 2016. Short-term prediction of vehicle waiting queue at ferry terminal based on machine
learning method. Journal of Marine Science and Technology 21 (4), 729–741.
Zhang, X., Peng, Y., Zhang, C., Wang, B., 2015. Are hybrid models integrated with data preprocessing techniques suitable for monthly
streamflow forecasting? Some experiment evidences. Journal of Hydrology 530, 137–152.
Zhang, Y., Wang, J., 2018. A distributed approach for wind power probabilistic forecasting considering spatio-temporal correlation
without directaccess to off-site information. IEEE Transactions on Power Systems 33 (5), 5714–5726.
Zhao, H.-X., Magoulès, F., 2012. A review on the prediction of building energy consumption. Renewable and Sustainable Energy
Reviews 16 (6), 3586–3592.
Zheng, J., Xu, C., Zhang, Z., Li, X., 2017. Electric load forecasting in smart grids using long-short-term-memory based recurrent
neural network. In: 2017 51st Annual Conference on Information Sciences and Systems (CISS). IEEE, pp. 1–6.
Zhou, C., Viswanathan, S., 2011. Comparison of a new bootstrapping method with parametric approaches for safety stock determi-
nation in service parts inventory systems. International Journal of Production Economics 133 (1), 481–485.
Zhou, L., Zhao, P., Wu, D., Cheng, C., Huang, H., 2018. Time series model for forecasting the number of new admission inpatients.
BMC Medical Informatics and Decision Making 18 (1), 39.
Zhou, Z., Matteson, D. S., 2016. Predicting Melbourne ambulance demand using kernel warping. The Annals of Applied Statistics
10 (4), 1977–1996.

240
Zhu, S., Dekker, R., van Jaarsveld, W., Renjie, R. W., Koning, A. J., 2017. An improved method for forecasting spare parts demand
using extreme value theory. European Journal of Operational Research 261 (1), 169–181.
Ziegel, J. F., Gneiting, T., 2014. Copula calibration. Electronic Journal of Statistics 8 (2), 2619–2638.
Ziel, F., Berk, K., 2019. Multivariate forecasting evaluation: On sensitive and strictly proper scoring rules. arXiv:1910.07325.
Ziel, F., Steinert, R., 2016. Electricity price forecasting using sale and purchase curves: The X-Model. Energy Economics 59, 435–454.
Ziel, F., Steinert, R., 2018. Probabilistic mid-and long-term electricity price forecasting. Renewable and Sustainable Energy Reviews
94, 251–266.
Ziel, F., Weron, R., 2018. Day-ahead electricity price forecasting with high-dimensional structures: Univariate vs. multivariate mod-
eling frameworks. Energy Economics 70, 396–420.
Zipf, G. K., 2016. Human Behavior and the Principle of Least Effort: An Introduction to Human Ecology. Ravenio Books.
Žmuk, B., Dumičić, K., Palić, I., 2018. Forecasting labour productivity in the European Union member states: Is labour productivity
changing as expected? Interdisciplinary Description of Complex Systems 16 (3-B), 504–523.
Zou, H., 2006. The adaptive Lasso and its oracle properties. Journal of the American Statistical Association 101:476, 1418–1429.
Zou, H., Hastie, T., 2005. Regularization and variable selection via the elastic net. Journal of the Royal Statistical Society, Series B
(Statistical Methodology) 67, 301–320.
Zou, H., Hastie, T., Tibshirani, R., 2006. Sparse principal component analysis. Journal of Computational and Graphical Statistics 15,
262–286.
Zwijnenburg, J., 2015. Revisions of quarterly GDP in selected OECD countries. OECD Statistics Briefing July 2015 - No. 22, 1–12.

241