Bhandari et al.

Financial Innovation (2024) 10:75 Financial Innovation

https://doi.org/10.1186/s40854-023-00604-0

RESEARCH Open Access

Implementation of deep learning models

in predicting ESG index volatility
Hum Nath Bhandari1, Nawa Raj Pokhrel2, Ramchandra Rimal3* , Keshab R. Dahal4 and Binod Rimal5

*Correspondence:
[email protected] Abstract
1
Department of Mathematics, The consideration of environmental, social, and governance (ESG) aspects has become
Roger Williams University, 1 Old an integral part of investment decisions for individual and institutional investors. Most
Ferry Rd., Bristol, RI 02809, USA recently, corporate leaders recognized the core value of the ESG framework in fulfill-
2
Computer Science and Physics
Department, Xavier University ing their environmental and social responsibility efforts. While stock market predic-
of Louisiana, 1 Drexel Dr, New tion is a complex and challenging task, several factors associated with developing
Orleans, LA 70125, USA an ESG framework further increase the complexity and volatility of ESG portfolios
3
Department of Mathematical
Sciences, Middle Tennessee compared with broad market indices. To address this challenge, we propose an inte-
State University, 1301 E Main St, grated computational framework to implement deep learning model architectures,
Murfreesboro, TN 37132, USA specifically long short-term memory (LSTM), gated recurrent unit, and convolutional
4
Department of Mathematics,
State University of New York neural network, to predict the volatility of the ESG index in an identical environment.
Cortland, 22 Graham Ave, A comprehensive analysis was performed to identify a balanced combination of input
Cortland, NY 13045, USA features from fundamental data, technical indicators, and macroeconomic factors
5
Department of Mathematics,
The University of Tampa, 401 W to delineate the cone of uncertainty in market volatility prediction. The performance
Kennedy Blvd, Tampa, FL 33606, of the constructed models was evaluated using standard assessment metrics. Rigorous
USA hyperparameter tuning and model-selection strategies were implemented to identify
the best model. Furthermore, a series of statistical analyses was conducted to validate
the robustness and reliability of the model. Experimental results showed that a single-
layer LSTM model with a relatively small number of neurons provides a superior fit
with high prediction accuracy relative to more complex models.
Keywords: ESG investing, ESG index, Deep learning, Machine learning, Volatility
prediction

Introduction
Traditional investors typically focus on investment returns in terms of profitability and
meticulously scrutinize financial reports to determine the best-performing stocks in the
market. A recent change in the mindsets of stakeholders and investors also considers the
non-financial impacts of investment decisions. Companies are evaluated under a broad
spectrum of environmental, social, and governmental (ESG) factors (Clementino and
Perkins 2021). Environmental factors mainly focus on natural resources such as energy
efficiency, biodiversity, pollution mitigation, water usage, and climate change. Similarly,
social components primarily cover the welfare of the society as a whole, including labor
standards, wages, benefits, affordable housing, education, workforce diversity, racial

© The Author(s) 2024. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits
use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original
author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third
party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the mate-
rial. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or
exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://
creativecommons.org/licenses/by/4.0/.
Bhandari et al. Financial Innovation (2024) 10:75 Page 2 of 24

justice, and health safety. Finally, the government strategically manages environmental
and social issues such as corporate board composition and overall structure, strategic
sustainability and oversight compliance, political contribution and lobbying, bribery,
and corruption.
ESG investing has gained tremendous popularity recently, as society expects com-
panies’ corporate and social responsibility efforts (Tucker and Jones 2020). To flourish
in the long run, financial institutions focus on the risk of investment and return on the
portfolio and evaluate whether the companies have embraced the agendas raised by
ESG. According to the Morningstar report, investors in the US poured a record $69.2
billion into ESG funds in 2021, three times higher than in 2020 (CNBCNews, June 5,
2022). US investors had access to more than 550 ESG related mutual funds and exchange
traded funds (ETFs) as of June 5, 2022, which is more than double in the past five years.
Similarly, in the European market, a total of $278 billion ESG related ETFs were under
management by 2021 (IR Magazine, Jan 18, 2022). From the perspectives of consum-
ers and investors, the global trend of sustainable investing is exponentially increasing.
Prominent industry leaders are beginning to acknowledge the importance of ESG by
providing the required information to ESG rating agencies, assuring ESG commitment,
and issuing sustainability reports to the public.
There is no consensus on a framework for ESG. Several indices and frameworks are
available in the market to better guide companies and inform investors. Some dominant
international frameworks include the Global Reporting Initiative (GRI) standards, Sus-
tainability Accounting Standards Board (SASB) standards, United Nations Principles for
Responsible Investment (UNPRI), and United Nations Sustainable Development Goals
(UNSDG). The scoring methodologies measure different parameters; thus, company
names may appear in one framework but not in the other. Controversies exist regarding
the agendas and their numerical quantification considered in every ESG framework. No
universally accepted framework, model, algorithm, or rule of thumb is available for solv-
ing this problem. However, it always helps stakeholders to delineate the cone of uncer-
tainty if human judgment and intuition are amalgamated over controversies based on
the context of the problem.
Although the idea of ESG-focused investing is relatively new, several high-profile
investment firms have begun to construct ESG indices by tracking companies commit-
ted to creating more environmentally friendly and sustainable business models. Some
of these include the S &P 500 ESG index, the Dow Jones Sustainability World Index,
MSCI World ESG Focus Index, and MSCI Emerging Markets ESG Focus Index. Several
mutual funds and ETFs provide investment opportunities for ESG savvy investors. These
include the Xtrackers S &P 500 ESG ETF (SNPE), SPDR S &P 500 ESG ETF (EFIV),
Invesco MSCI Sustainable Future ETF (ERTH), iShares MSCI Global Sustainable Devel-
opment Goals ETF (SDG), Fidelity International Sustainability Index Fund (FNIDX), and
Vanguard FTSE Social Index Fund (VFTAX).
Stock market prediction is a complex and challenging task because of its nonpara-
metric, nonlinear, and chaotic behavior (Ahangar et al. 2010). In addition, investment
decisions are not always made simply by looking at structural data, such as balance
sheets, financial report cards, company valuations, and volumes of shares traded in a
specific range. Investors go beyond these factors and consider whether a company has
Bhandari et al. Financial Innovation (2024) 10:75 Page 3 of 24

incorporated ESG agendas into its business models. These factors depend mainly on the
nature of the company and its associated market structure from local and global per-
spectives. Consequently, ESG factors exhibit additional complexity in an already com-
plex and volatile market. Thus, there is a pressing demand to develop a proper model
that helps measure the performance and volatility of ESG indices to minimize related
risks and better inform stakeholders before making responsible financial decisions.
Most classical time series models assume linear data relationships. However, this
assumption raises significant concerns regarding the robustness of these classical mod-
els when applied to real-world time series data that frequently exhibit nonlinear behav-
ior. Moreover, classical machine learning approaches struggle to capture long-term
dependencies within time series data. This is where deep learning models have come to
the forefront, as they effectively address these limitations. Deep learning excels at com-
prehending intricate patterns and connections within financial data, offering benefits
such as automated feature extraction, nonlinear handling, temporal dependency cap-
ture, adaptability to changing conditions, and efficient management of extensive data-
sets. These attributes collectively position deep learning models as superior tools for
precisely predicting ESG index volatility compared to conventional models.
Many studies have been conducted to build efficient predictive models using machine
learning and deep learning techniques (Nabipour et al. 2020; Wang et al. 2020; Sen and
Chaudhuri 2018). Some of these studies focus on predicting the price and/or volatility of
ESG related indices (Guo et al. 2020; Lee et al. 2022; Raman et al. 2020). Varying degrees
of success were observed, based on the accuracy and robustness of the models. The most
widely used deep learning architectures are long short-term memory (LSTM), convolu-
tional neural networks (CNN), gated recurrent units (GRU), and their respective hybrid-
ization techniques (Lin and Jin 2023)
We noticed several gaps in the literature. For instance, researchers often utilize the
stated methods to speak of oneself with pride in terms of a model’s accuracy. However,
the model framework, underlying assumptions, and implementation differ. Thus, it is
difficult to perform an unbiased comparison between published research articles, even
if they use the same deep learning architecture to construct their predictive models. Fur-
thermore, the authors could not find a transparent and data-driven approach for fine-
tuning the model hyperparameters. In addition, several previous studies have focused
on price prediction rather than volatility prediction, which is the focus of this study. This
trend can be attributed to ESG savvy investors’ concerns with the volatility and risks
associated with their investment portfolios, rather than short-term returns. In addition,
there is a significant lack of analysis on the robustness of the constructed models.
The current study aimed to fill these gaps by (a) providing an integrated computational
framework to implement deep learning model architectures to predict the volatility of
the ESG index in an identical environment; (b) gathering multifaceted information that
directly and indirectly affects the ESG index, putting them together to construct a well-
balanced set of input features; (c) implementing an extensive and data-driven approach
for hyperparameter tuning and model selection; and (d) conducting statistical analyses
to validate and verify the reliability and robustness of the model.
A complete roadmap for achieving this goal is presented in the schematic diagram in
Fig. 1. Well-balanced input features were incorporated into the spheres of fundamental
Bhandari et al. Financial Innovation (2024) 10:75 Page 4 of 24

Fig. 1 Schematic diagram of the proposed research framework

data, macroeconomic data, and technical indicators. The collected data were normalized
using the min-max technique, and input sequences for the models were created using a
specific time step. Hyperparameters such as the number of neurons (or filters), epochs,
learning rate, and batch size were tuned using regularization techniques to optimize the
model performance. Once the hyperparameters were tuned, the models were trained to
predict the volatility of the ESG index. Finally, the model quality was assessed using the
RMSE, MAPE, and R-scores of a test set.
The remainder of this paper is organized as follows. Section explains the related work
in this field. The data collection and feature selection procedure is explained in Sect. .
Modeling approaches are discussed in Sect. . Section discusses the experiment and
results, followed by discussion in Sect. . Section discusses the ethics and implications.
Finally, Sect. presents the conclusions and future work, followed by acknowledgments
and a list of references.

Related work
Although ESG investing is a relatively new thematic investment idea yet to be fully
adopted by the mainstream investment community, various studies have been conducted
to understand the importance of ESG criteria in portfolio construction and optimiza-
tion, the integration of ESG factors in machine learning models for price and volatility
predictions, and the role of ESG factors during systemic crises.
Some researchers have explored the importance of ESG factors in portfolio construc-
tion and optimization. Vo et al. developed a deep responsible investment portfolio to
predict quarterly and yearly stock returns, which they then combined with ESG ratings
in their modified mean-variance ESG model to construct and rebalance socially respon-
sible investment (SRI) portfolios (Vo et al. 2019).
Xidonas and Essner employed a minimax optimization approach to enhance portfo-
lio optimization, which entailed the integration of key ESG risk performance factors.
The minimax methodology facilitates the optimization of individual security weights
within the portfolio, aiming to reduce deviations from ESG targets. This is achieved by
simultaneously minimizing the maximum risks and maximizing the attainment of ESG
investment objectives. They tested the models’ performance on multiple European and
American stock indices and demonstrated better risk-adjusted returns than the bench-
marks (Xidonas and Essner 2022). Berg et al. conducted an empirical analysis of ESG
investments that quantified the returns associated with ESG investment strategies and
assessed their financial performance. This study analyzed a diverse set of companies
and industries to evaluate the impact of ESG metrics on investment outcomes. These
findings highlight the correlation between ESG scores and stock returns and indicate
a potential link between sustainable practices and financial success (Berg et al. 2023).
Lucia et al. conducted a case study to explore whether ESG practices led to better
financial performances in 1038 public enterprises in Europe. Their findings suggest a
Bhandari et al. Financial Innovation (2024) 10:75 Page 5 of 24

relationship between ESG variables and improved financial performance (De Lucia et al.
2020). Hang and Chen proposed two SRI portfolio construction models, namely double-
screening socially responsible investments I and II, which utilized a double-screening
mechanism and an extreme learning machine model with genetic algorithm optimiza-
tion to predict stocks and integrate ESG factors to determine the investment propor-
tion of the screened stocks. The study claimed that the proposed models exhibited better
performance (Zhang and Chen 2011). Umar et al. investigated the relationship between
the cryptocurrency environmental attention index and the volatility and return on assets
categorized as either green or dirty (Umar et al. 2022). Their findings suggest that dirty
equities and bonds are the main drivers of return spillover, while dirty equities transmit
volatility spillover, and that environmental attention has a greater effect on equities than
on bonds. These findings provide insights into investment, hedging, and policymaking
decisions as well as the potential usefulness of ESG investments in providing diversifica-
tion. All of the above studies support the idea that ESG has a positive impact on portfo-
lio construction and optimization.
Efforts have been made to integrate ESG factors into machine learning techniques to
enhance the accuracy of stock price predictions by identifying the underlying ESG alpha.
For instance, Chen et al. utilized ESG scholar data to establish an automatic trading
strategy and proposed a practical machine learning approach to quantify a company’s
ESG premium and capture ESG alpha (Chen and Liu 2020). Their study involved cre-
ating an ESG investment universe, conducting feature engineering on the ESG scholar
data of companies, and training the proposed models using financial indicators and ESG
scholar data. They used an ensemble method to forecast stock prices and provided rec-
ommendations for portfolio construction, trading, and rebalancing. According to this
study, the proposed ESG alpha strategy generated impressive cumulative returns from
the proposed portfolio compared with several benchmarks. Similarly, Magrot et al.
designed and implemented a machine learning algorithm capable of identifying patterns
between ESG profiles and performance (Margot et al. 2021). Their algorithm generates
a set of rules, each of which identifies a region in the high-dimensional space of ESG
features in which excess stock returns can be predicted. This study empirically demon-
strates the correlation between ESG profiles and financial performance.
ESG investors are typically savvy investors who prioritize the volatility and risk associ-
ated with their investment portfolios over short-term returns. A few researchers have
focused on incorporating ESG criteria into the development of efficient volatility pre-
diction models. For example, Sabbaghi conducted empirical investigations of asym-
metric volatility in ESG investing using Morgan Stanley Capital International (MSCI)
indices and found that the impact of news on the volatility of ESG firms is greater for
bad news than for good news (Sabbaghi 2020). Additionally, the impact of bad news on
the volatility of ESG firms is smaller for small-cap ESG firms than for large- and mid-cap
ESG firms. By contrast, Guo et al. implemented a new deep learning framework called
ESG2Risk to predict the future volatility of stock prices using ESG news (Guo et al.
2020). The study concluded that ESG news has a significant impact on the future returns
and risks of companies and can therefore be considered a relevant factor when mak-
ing investment decisions. The studies discussed above, including (Yu et al. 2022; Daniali
Bhandari et al. Financial Innovation (2024) 10:75 Page 6 of 24

et al. 2021), provide evidence that machine learning models that incorporate ESG factors
outperform other models in predicting volatility.
Market volatility increases during systemic crises, such as recessions, pandemics, and
wars. The inclusion of specific factors in the model is required to capture these effects.
Umar et al. investigated how social media coverage of the Covid-19 pandemic affected
ESG leader indices in different regions, identifying periods of low, medium, and high
coherence between the media coverage index and the price movements of the ESG
leader indices Umar and Gubareva (2021). The periods of low coherence suggest that
ESG investments could potentially provide diversification benefits during a systemic
pandemic like Covid-19. Moreover, Akhtaruzzaman et al. found that media coverage
contributed to the spread of the contagion in both advanced and emerging equity mar-
kets, with the US being the most severely impacted country (Akhtaruzzaman et al. 2022).
Albuquerque et al. investigated the mechanism by which corporate social responsibility
(CSR) and ESG policies affect firms’ systematic risk by assuming CSR as a product dif-
ferentiation strategy. They claim that strong ESG firms face relatively less price-elastic
demand, which results in lower systematic risk due to a product differentiation strategy.
They concluded that consumers play a vital role in influencing firm policies and risk pro-
files (Albuquerque et al. 2019).
In summary, limited research has been conducted on ESG-related stock market port-
folios and volatility predictions compared to the volatility predictions of broad stock
market indices (Cho and Lee 2022; Koo and Kim 2023; Mittnik et al. 2015; Lu et al.
2022). The reviewed studies made significant contributions to integrating ESG into port-
folio construction, optimization, performance analysis, and risk assessment. However,
some of these studies focused solely on building a complex model, whereas others imple-
mented machine learning models without serious consideration of feature selection. An
efficient model is required that utilizes a balanced combination of input features, while
maintaining the simplicity of its architecture. Our study aimed to address these issues
by developing an integrated framework for implementing state-of-the-art deep learn-
ing models trained with the best possible set of influencing factors. The main goal was
to ensure a comprehensive understanding of the behavior of ESG investment portfolios
from multiple dimensions and offer valuable insights for future research.

Data description and preparation

This study used the S &P 500 ESG index, a popular ESG focused index in the US. It is a
broad-based market-cap-weighted index designed to measure the performance of secu-
rities meeting sustainability criteria while maintaining similar industry group weights to
the S &P 500 (Winegarden 2019; Gary 2019). S &P Global maintains the index under the
Dow Jones Indices (Indices 2016). The launch date of the index was January 28, 2019,
and the backward data assumption date was May 3, 2010. Factors such as fundamental
data, technical indicators, and macroeconomic variables may contribute directly or indi-
rectly to index value fluctuations (Serfling and Miljkovic 2011; Tien et al. 2021). The core
intrinsic fundamental data are extracted directly from the underlying index. Technical
indicators are byproducts of fundamental data that utilize standard mathematical equa-
tions to produce final numerical values. Macroeconomic variables were selected based
on their potential impact on the overall economy and broader markets.
Bhandari et al. Financial Innovation (2024) 10:75 Page 7 of 24

Input features, such as fundamental data and technical indicators, provide crucial
internal information about the overall quality of the underlying stocks as well as sup-
ply and demand situations in a given market environment. Other factors, namely mac-
roeconomic variables, contribute by providing information about the potential external
influence on the given index fluctuations, capturing the status of the overall economy
and broader markets. The incorporation of these comprehensive data sources is pivotal
for enhancing the predictive ability of the deep learning framework and ensuring a more
robust and accurate analysis of the complex dynamics of stock markets. Consequently,
insights gained from this holistic approach can significantly contribute to informed deci-
sion-making and more effective predictions.
The selected timeframe for the data was from 01–02-2013 to 12–30-2021, which
incorporates a major bear market during the COVID-19 pandemic in 2020. Thus, the
construction of the model, which includes both bear and bull markets, resembles the
overall market scenario.
S &P 500 ESG index is constructed primarily from the popular US broad market index.
Based on a thorough investigation of the related literature and also from the exploratory
data analysis, we can identify the following evidence.

• Finding 1: The information presented in Table 1 and Fig. 2 vividly reveal the fact that
the two indices are not identical in terms of their constituents and sector exposures
(Indices 2016).
• Finding 2: S &P 500 and S &P 500 ESG have almost similar patterns in terms of daily
returns and cumulative returns, as demonstrated in Fig. 3.
• Finding 3: Fig. 4 shows similar annualized rolling volatility and Sharpe ratio patterns
of these two indices in the given time interval. S &P 500 ESG index’s annualized
return is slightly higher than the S &P 500, but these higher returns come with higher
risks, as illustrated in Fig. 5.

Table 1 Top 10 constituents of S &P 500 and S &P 500 ESG as of May 31, 2022
Rank S &P 500 S &P 500 ESG
Company Sector Company Sector

1 Apple Inc. (AAPL) Information Technology Apple Inc. (AAPL) Information Technology
2 Microsoft Corp (MSFT) Information Technology Microsoft Corp (MSFT) Information Technology
3 Amazon.com Inc. Consumer Discretionary Amazon.com Inc. Consumer Discretionary
(AMZN) (AMZN)
4 Alphabet Inc A (GOOGL) Communication Alphabet Inc A (GOOGL) Communication Services
Services
5 Tesla Inc (TSLA) Consumer Discretionary Alphabet Inc C (GOOG) Communication Services
6 Alphabet Inc C (GOOG) Communication Unitedhealth Group Inc Health Care
Services ((UNH)
7 Berkshire Hathaway B Financials Nvidia Corp (NVDA) Information Technology
(BRK.B)
8 Johnson & Johnson (JNJ) Health Care Exxon Mobil Corp (XOM) Energy
9 Unitedhealth Group Inc Health Care JP Morgan Chase & Co Financials
(UNH) (JPM)
10 Nvidia Corp (NVDA) Information Technology Procter and Gamble (PG) Consumer Staples
Bhandari et al. Financial Innovation (2024) 10:75 Page 8 of 24

Fig. 2 Sector-wise composition of S &P 500 and S &P 500 ESG indices as of May 31, 2022

Fig. 3 Comparison of daily and cumulative returns of S &P 500 and S &P 500 ESG indices

Fig. 4 Rolling volatility and Sharpe ratio of S &P 500 and S &P 500 ESG indices

• Finding 4: The broad market macroeconomic features such as CBOE Volatility

Index, Interest Rate, and US Dollar Index have a similar impact on both indices, as
shown in Fig. 6. The data entries of the correlation matrix of broad market mac-
roeconomic features to the closing price and volatility of both S &P 500 and S &P
500 ESG indices show similar correlation.
Bhandari et al. Financial Innovation (2024) 10:75 Page 9 of 24

Fig. 5 Comparison of annualized returns and volatility(Left: Annualized returns and Right: Annualized
volatility)

Fig. 6 Correlation heatmaps (Left: S &P 500 data, Right: S &P 500 ESG data)

From the aforementioned evidence, we conclude that the S &P 500 ESG index cap-
tures broad US financial market behavior and exhibits similar functionality to the
S &P 500 index in terms of returns and volatility, irrespective of variations in their
constituents and sector exposures. Therefore, the features, particularly the mac-
roeconomic factors, used to predict the S &P 500 index (Bhandari et al. 2022a) can

Table 2 List of potential features for the model

Data Source Frequency Abbreviation

Fundamental
Close price S &P Dow Jones indices Daily ···
Macroeconomic
Cboe volatility index Yahoo Daily VIX
Interest rate FRED Daily EFFR
Civilian unemployment rate FRED Monthly UNRATE
Consumer sentiment index FRED Monthly UMCSENT
US dollar index Yahoo Daily USDX
Technical indicator
Volatility ··· Rolling ···
Moving average convergence divergence · · · Daily MACD
Relative strength index ··· Daily RSI
Sharpe ratio ··· Daily SR
Bhandari et al. Financial Innovation (2024) 10:75 Page 10 of 24

also be used in the S &P 500 ESG index. The complete input variables used in this
study are listed in Table 2, and their short descriptions are presented in the following
subsection.

Fundamental data
The first set of variables presented in Table 2 comprises fundamental or historical data
that provide basic information regarding the performance of the index. The closing price
is the final price of the index on a given trading day.

Macroeconomic data
The second set of variables shown in Table 2 comprises macroeconomic data that sig-
nificantly influence stock market performance by reporting the overall health of the
financial market (Bhandari et al. 2022a; Bhandari et al. 2022). We choose the CBOE
volatility index (VIX), interest rate (EFFR), civilian unemployment rate (UNRATE), con-
sumer sentiment index (UMCSENT), and US dollar index (USDX) as macroeconomic
factors (Chandra and Thenmozhi 2015; Ruan 2018; Bernanke and Kuttner 2005; Farsio
and Fazel 2013; Bock 2018; Baker and Wurgler 2007). These variables are representative
features that explain the overall status of the economy in the proposed model.

Technical indicators
The third set of variables, shown in Table 2, are technical indicators, including volatility,
moving average convergence divergence (MACD), relative strength index (RSI), and the
Sharpe ratio (SR). Volatility was used as both the input and response variables in this
study. First, monthly volatility is calculated as the rolling standard deviation of monthly
returns (21 trading days on average, based on the US market). Monthly volatility is then
√
annualized by multiplying it by 12:
Active traders use them extensively in the market because they are primarily designed
to analyze short-term price movements and are included in this study (Rodríguez-
González et al. 2011; Wilder 1978; Anghel 2015; Chong et al. 2014; Chong and Ng 2008;
Eric et al. 2009; Murphy 1999; Wang and Kim 2018; Schmidt 2022; Goyal and Aggarwal
2014).

Modelling approach
Deep learning models: LSTM, GRU, and CNN
Let (xt , yt ) be a input–output pair of the model, where xt ∈ Rk×1 is the input feature,
and yt ∈ R is the output at times t = 1, 2, . . . , n. Here, k and n are the number of input
features and total number of observations, respectively. Furthermore, to incorporate
the time step into LSTM, GRU, and CNN architectures, the input sequence Xt was cre-
ated by taking m continuous sequence xt : xt+m−1, which is a matrix of shape k × m for
t ∈ {1, 2, . . . , n − m − 1}.
LSTM is a recurrent neural network consisting of an input, hidden state, cell state,
and output. It is designed using a gate mechanism (Hochreiter and Schmidhuber 1997;
Gers et al. 2000, 2003). LSTM has four gates: input, update, forget, and output, as shown
Bhandari et al. Financial Innovation (2024) 10:75 Page 11 of 24

Fig. 7 Long short-term memory(LSTM) architecture (Bhandari et al. 2022a)

in Fig. 7 (Bhandari et al. 2022a). At time t, the gates and layers compute the following
functions:

it = σ (Wi xt + Whi ht−1 + bi ),

ft = σ (Wf xt + Whf ht−1 + bf ),
ot = σ (Wo xt + Who ht−1 + bo ),
c˜t = tanh(Wc xt + Whc ht−1 + bc ),
ct = ft ⊗ ct−1 + it ⊗ c˜t ,
ht = ot ⊗ tanh(ct )

where σ and tanh represent the sigmoid and hyperbolic tangent functions, respectively,
the operator ⊗ is the element-wise product, W ∈ Rd×k , Wh ∈ Rd×d are the weight
matrices, and b ∈ Rd×1 is the bias vector. Moreover, d denotes the hidden size (Greff
et al. 2017; Qiu et al. 2020; Lei et al. 2019).
The input gate identifies information that must be updated from the change gate.
The output of the forget gate is between 0 and 1 through a sigmoid activation func-
tion. This identifies the information required to forget former cell state ct−1. It stores
all the information in the cell if the output is 1. However, it forgets all the information
from the previous cell state if the output is 0. The output gate determines which infor-
mation is to be taken as the output from the present cell state, and the output (ht , ct )
of LSTM is a feature representation of the input sequence Xt at time t, which can be
expressed as follows:

(ht , ct ) = LSTM(Xt , ht−1 , ct−1 , w).

GRU is a simplified version of LSTM (Chollet 2017). The short-term (ht ) and long-term
(ct ) information of LSTM are merged into a single vector ht in GRU. In contrast to the
four gates in LSTM, GRU has three gates: reset gate, change gate, and update gate, as
shown in Fig. 8. The update gate of GRU is equivalent to the forget gate and input gate
Bhandari et al. Financial Innovation (2024) 10:75 Page 12 of 24

Reset Gate

*
+ * + =

Change Gate

*
+ + =

Update Gate

*
+ * + =

Fig. 8 Gated Recurrent Unit (GRU) architecture (Pokhrel et al. 2022)

of LSTM (Gáeron 2019). Thus, a single gate decides what to forget and update in GRU
instead of the two gates in LSTM.
At time t, the gates and layers compute the following functions:

ut = σ (Wz xt + Whz ht−1 + bu ),

rt = σ (Wr xt + Whr ht−1 + br ),
h̃t = tanh(Wc xt + Whc (rt ⊗ ht−1 ) + bc ),
ht = (1 − ut ) ⊗ ht−1 + ut ⊗ h̃

The output ht of GRU is a feature representation of the input sequence Xt at time t and is
calculated as follows:

ht = GRU (Xt , ht−1 , w).

The CNN architecture has the following components: input, convolutional layer with a
nonlinear activation function, a pooling layer, a fully connected layer, and an output. All
the layers in a CNN have training parameters, except for the pooling layer. A CNN views
a time step as a sequence in which convolutional operations can be performed on a one-
dimensional image. Because each series contains observations at the same time step, the
input time series is parallel. We can reconfigure these three data arrays (no. of samples,
time steps, and no. of features) as a single dataset, where each row is a time step, and
each column is a separate time series (Brownlee 2018b, c). We have n − Ts many matri-
ces of size Ts × k as in LSTM and GRU, and each matrix is treated as an image of size
k × Ts in the CNN. The output ht of the CNN is a feature representation of the input
sequence Xt at time t, which can be expressed as

ht = CNN (Xt , w).

For each image, we use m filters and slide each filter on the time axis with a stride of
one. Then, after the convolution operation, we obtain m feature maps from m filters.
Bhandari et al. Financial Innovation (2024) 10:75 Page 13 of 24

Fig. 9 CNN architecture with m filters for multivariate time series prediction (Pokhrel et al. 2022; Rimal 2022)

After the convolution operations, we use nonlinear activation functions such as ReLU
or Leaky ReLU. A pooling operation is performed for downsampling. Subsequently,
the feature maps from each filter are vectorized into a single sequence to form a fully
connected layer. Finally, the output yˆ1 is predicted using a linear activation function, as
shown in Fig. 9.

Experimental design and results

The primary goal of this study was to conduct a comparative analysis of the perfor-
mance of LSTM, GRU, and CNN models in volatility prediction. Figure 10 shows the
original time series of the annualized rolling volatility of the S &P500 ESG index for the

Fig. 10 S &P 500 ESG annualized rolling volatility

Bhandari et al. Financial Innovation (2024) 10:75 Page 14 of 24

Fig. 11 Experimental design

01–02-2013 textemdash 12–30-2021 interval, which exhibits complex, noisy, and vola-
tile behavior.
To achieve the stated goal, as shown in Fig. 11, the overall experiment was divided into
five phases: (a) environmental setup and input preparation, (b) model construction and
hyperparameter tuning, (c) identifying the best-performing models from the respective
architectures, (d) identifying the overall best-performing model, and (e) performing sta-
tistical analysis.

Environmental setup and input preparation

Table 3 summarizes the computational framework of the experiments. The experi-
ments used the Python programming environment and TensorFlow and Keras APIs. The
machine configuration and architecture used in the experiments are also listed in the
Table 3
As part of the input/output preparation, the original dataset was first divided into
training and test sets at a ratio of 4:1. Among the training data, 25% was separated for
validation, which accounted for 20% of the total data. A validation set was used for
hyperparameter tuning. After obtaining the optimal hyperparameters, the validation
data were added to the training set. The overall distribution of the data is presented in
Table 4.

Table 3 Computing environmental setup

Machine configuration Google Colab with NVIDIA-SMI 495.44 GPU

Environment Python 3.6.0, TensorFlow, and Keras APIs
Architecture LSTM, GRU, CNN

Table 4 Overall distribution of training, validation, and test data

Data Dates No. of samples

Complete data 2013–01–02 to 2021–12–30 2264

Training 2013–01–02 to 2020–03–16 1811
Validation 2018–05–29 to 2020–03–16 453
Test 2020–03–17 to 2021–12–30 453
Bhandari et al. Financial Innovation (2024) 10:75 Page 15 of 24

Because the range of values for the input features varied widely, a min–max nor-
malization technique was implemented. The normalized data were in the form of a 2D
array (number of observations and features). However, the proposed model architecture
requires 3D input data. Thus, it was converted into a 3D array (number of observations,
time steps, and number of features) by incorporating the time step before being fed into
the model. The prediction accuracy of the constructed model was assessed using three
performance metrics: RMSE, MAPE, and R. The stated matrices help determine the best
model in terms of accuracy and reliability.

Model construction and hyperparameter tuning

We constructed deep learning models, each of which consisted of an input layer, an
LSTM/GRU/CNN layer, and a dense output layer with linear activation. Early stopping
criteria were implemented to address the consequences of underfitting and overfit-
ting that can occur when training neural networks. This approach allowed us to spec-
ify a large number of epochs and stop training when the model’s performance stopped
improving on the validation data (Brownlee 2018a).
After constructing the model, we performed a hyperparameter tuning process in
which each model identified its best set of hyperparameters from multiple avenues. This
included three different optimizers (Adam, Adagrad, and Nadam), three different learn-
ing rates (0.1, 0.01, and 0.001), and three batch-size options (4, 8, and 16). Therefore,
3 × 3 × 3 = 27 possible choices were available for each model for identifying the best
combination. We performed ten independent replicates for each model before calculat-
ing the average scores to address the model’s stochastic behavior. The best model was
selected based on the lowest possible average RMSE score calculated on the validation
dataset. Thus, we executed three architectures—(LSTM, GRU, and CNN) × six models
for each architecture (number of different neurons) × 27 (possible combinations for each
model) = 486 instances—during the complete hyperparameter tuning process. The opti-
mal set of hyperparameters for each model architecture is presented in Table 5.

Identifying the best performing models from respective architectures

Once the hyperparameter tuning process was completed, the models were set with their
corresponding hyperparameters. Finally, all models (6 ∗ 3 = 18) were trained in full
scale with the best hyperparameters. Fully trained models were implemented on the test

Table 5 Optimal hyperparameters for LSTM, GRU, and CNN models

No. of LSTM GRU​ CNN
neurons/
filters Optimizer Learning Batch Optimizer Learning Batch Optimizer Learning Batch
rate size rate size rate size

10 Adam 0.001 8 Adagrad 0.1 16 Adagrad 0.01 8

30 Adagrad 0.1 4 Adagrad 0.01 8 Adagrad 0.01 8
50 Adagrad 0.01 16 Adagrad 0.01 16 Adagrad 0.01 4
100 Adagrad 0.001 4 Adagrad 0.001 16 Adagrad 0.01 4
150 Adagrad 0.001 4 Adagrad 0.001 16 Adagrad 0.001 4
200 Adagrad 0.001 16 Adagrad 0.001 16 Adagrad 0.001 16
Bhandari et al. Financial Innovation (2024) 10:75 Page 16 of 24

Fig. 12 Average scores obtained from LSTM, GRU, and CNN models: a RMSE, b MAPE, and c R on test dataset

Fig. 13 Boxplots of evaluation metrics for a LSTM models, b GRU Models, and c CNN Models

data to verify their performance and reliability. We replicated each model 30 times to
address the stochastic behavior of the deep learning models. Figure 12 shows a graphi-
cal representation of the average scores produced by the employed model architectures
(LSTM, GRU, and CNN). The subplots (a), (b), and (c) show the overall patterns of the
average RMSE, MAPE, and R-scores for each model architecture.
Observing the performance scores in a holistic approach, for LSTM, the average
RMSE and MAPE scores were low with 10 neurons. Thereafter, no significant decreas-
ing trend appeared. Similarly, the highest average R-score was observed for 10 neurons.
In addition, GRU with 50 neurons provided the smallest average RMSE and MAPE, and
the most significant average R score. The CNN model with 100 neurons had the smallest
Bhandari et al. Financial Innovation (2024) 10:75 Page 17 of 24

average RMSE and MAPE and the largest R score. Furthermore, the distributions of the
RMSE, MAPE, and R scores and their variabilities obtained from the 30 replicates are
presented in Figs. 12 and 13.
Based on the comparative analysis, it can be concluded that the 10 neurons LSTM, 50
neurons GRU, and 100 neurons CNN were the best in their respective categories. The
list of best-performing models from the respective architectures, along with their opti-
mal hyperparameters, is highlighted in Table 5 using bold letters.

Identifying overall best model

After identifying the best models from the respective architectures, we compared the
performance scores to identify the best model among the three. Table 6 presents the sta-
tistics of the performance scores obtained from the three best models. The LSTM with
10 neurons showed the smallest RMSE (0.5849), MAPE (0.1425), and R (0.9952) scores.
The GRU with 50 neurons had the second-smallest average RMSE (0.7621) and MAPE
(0.2046), and the second-largest R-score (0.9917). Similarly, the standard deviation of the
R scores was the smallest and the standard deviations of RMSE and MAPE scores were
slightly larger for the best-performing LSTM model compared with those of the best-
performing GRU model. In addition, Fig. 13 illustrates that the overall distributions of
the scores were approximately symmetric with relatively small variability, indicating the
consistent performance of the three best-performing models. Thus, Table 6 and the dis-
tribution observed in Fig. 13 suggest that the LSTM model with 10 neurons is the win-
ner, followed by GRU with 50 neurons and CNN with 100 neurons.
Figure 14 shows the true vs. predicted plots that gauge the goodness of fit to determine
the quality of the prediction obtained from the training and test data. The blue dots rep-
resent the actual versus predicted values, and the olive dotted line shows the best fit of
each plot ( y = x ). The overall fit of the training data is almost indistinguishable in all
three subplots of Fig. 14a, despite the relatively better performance of LSTM. In the test
data, the predicted values deviated to a greater extent from the actual values compared
with the training data, as expected. Among the three subplots in Fig. 14b, LSTM shows a
superior fit compared with GRU and CNN.
Figure 15 shows the actual time series together with the predicted volatility obtained
from the three best models. The blue curves represent the actual values, whereas the

Table 6 Performance scores of the models on the test data

Models Metrics → RMSE MAPE R

LSTM Mean ± Std 0.5849 ± 0.1136 0.1425 ± 0.0372 0.9952 ± 0.0008

Minimum 0.40172 0.0917 0.9926
Maximum 0.8939 0.2661 0.9964
GRU​ Mean ± Std 0.7621± 0.0855 0.2046 ± 0.0247 0.9917 ± 0.001354
Minimum 0.6039 0.1447 0.9881
Maximum 0.9346 0.246957 0.9938
CNN Mean ± Std 1.3661048 ± 0.2512 0.37198 ± 0.0935 0.9679 ± 0.0252
Minimum 0.9666 0.2120 0.8427
Maximum 2.3382 0.6286 0.9856
The bold values represent the scores associated with the best performing model on the test data
Bhandari et al. Financial Innovation (2024) 10:75 Page 18 of 24

Fig. 14 True versus predicted value plots of the best performing LSTM, GRU, and CNN models

Fig. 15 Time series plots of the true and predicted values obtained from three best performing models

maroon and olive curves represent the values predicted from the training and test data,
respectively. As shown in the subplots in Fig. 15a and b, the prediction curve obtained
from the LSTM model captures the fluctuations more accurately in almost every situa-
tion. However, the GRU and CNN struggle to capture actual values, particularly in the
test data. It is clear that the LSTM provided a superior fit compared with the others.
Bhandari et al. Financial Innovation (2024) 10:75 Page 19 of 24

Table 7 Test statistics and p-values from normality test of RMSEs of the models
LSTM GRU​ CNN

Test statistics 4.20 0.5329 0.7896

P-value 0.1223 0.766 0.673

Table 8 Test statistics and p-values from two samples t-test for pairwise comparison of model
performance
(LSTM, GRU) (LSTM, CNN) (GRU, CNN)

Test statistics − 6.7063 − 18.8414 − 15.3499

P-value 1.2476 × 10−8 1.2305 × 10−23 2.3429 × 10−18

Statistical analysis
To validate the reliability of the model outcome, we conducted a statistical analysis
to identify whether the performances of the three best models differed significantly.
We performed a pairwise comparison of the mean RMSEs of the three models using
Welch’s two-sample t-tests. The normality test of RMSEs based on D’Agostino and
Pearson (D’agostino and Pearson 1973) ensures that the RMSEs of the three models
follow normal distributions, as the p-values are significantly higher than the signifi-
cance level α = 0.05 as presented in Table 7.
The test statistics and p-values from the two-sample t-test are listed in Table 8. A
significant difference exists between the mean RMSEs of the pairs (LSTM, GRU),
(LSTM, CNN), and (GRU, CNN). The pairwise model comparison produced an out-
come in favor of the LSTM model. Hence, we conclude that the LSTM model with
10 neurons best predicts the volatility of the S &P500 ESG index.

Discussion
This study developed an efficient model for predicting the volatility of the broader
ESG index of the stock market using deep learning architectures, such as LSTM,
GRU, and CNN. This study utilized a diverse set of features from multiple avenues
that contribute to ESG index volatility and compared the model performance. The
researchers collected data from various sources and prepared the data for modeling.
The study rigorously followed standard guidelines for predictive modeling and iden-
tified the overall best model with the best fit and highest prediction accuracy. The
models were trained using data from both bull and bear market conditions, includ-
ing the great recession of 2007–2009 and the COVID-19 market downturn, and their
performances were evaluated using several measures. Thus, the developed model
can make reasonable predictions, even in highly volatile market situations.
The research can be extended to model unusual volatility during a systemic cri-
sis, which may require a close attention to the specific crisis, and identify additional
features that can influence investors sentiment during that crisis. Some studies have
discussed the importance of studying this scenario and suggested that studying the
Bhandari et al. Financial Innovation (2024) 10:75 Page 20 of 24

performance of equities during a systemic crisis requires special treatment because

several unusual factors contribute to volatility (Jabeur et al. 2021; Kou et al. 2019;
Chatzis et al. 2018; Lee et al. 2019; Engelhardt et al. 2021). Another potential exten-
sion could be to utilize the predictive power of the proposed model for investment
portfolio construction, optimization, and analyzing risk-adjusted returns. Recent
studies in this demanding research area include the development of automatic clus-
tering and fuzzy system-based approaches to optimize investment portfolios by ana-
lyzing large-scale financial data (Li et al. 2021; Kou et al. 2021).

Ethics and implications

The model development process is not driven by profit maximization. All major ethi-
cal attributes, such as transparency, integrity, and candor, are internalized to maintain
the trustworthiness of the stakeholders. This study used a publicly available dataset
without manipulation. Machine learning scripts are completely inspected, inter-
pretability of the final outcome concerning domain knowledge is not sacrificed. The
reported performance of the model is the average performance of the out-of-sample
data based on several replications. Thus, the results can be used as additional infor-
mation to make an investment decision that upholds investors’ confidence. However,
investment decisions should not rely entirely on the research outcomes. Investors are
expected to perform due diligence and consider their risk tolerance under various
market conditions. A reasonable forecast depends not only on the outcome of the
specific model but also on the volatile nature of the stock market, especially during
geopolitical tension, global supply chain disturbances, war, pandemics, and various
other market risks. Thus, stakeholders can benefit if the market’s current behavior is
appropriately analyzed and amalgamated with the model’s outcome.
Equity traders, individual investors, and portfolio managers intrinsically want to
predict volatility using projected risks. This study demonstrates the potential of a
neural network architecture to delineate the cone of uncertainty in market volatility
prediction. Moreover, academic researchers can build the proposed model framework
to expand horizons in the field of sequential data modeling.

Conclusion
Predicting the volatility of the stock market is of great interest to finance practitioners
to best allocate their assets and academics to build an optimal model for consistent
predictions with a high level of accuracy. Predicting a volatile market is challeng-
ing because of its noisy and nonlinear behavior. Multifaceted factors, both local and
global, may directly or indirectly affect predictions. This study built predictive models
using 10 predictors that fall under fundamental, macroeconomic, and technical data.
A comparative analysis of S &P500 ESG index volatility prediction was performed
using deep learning architectures, namely LSTM, GRU, and CNN. An extensive data-
driven approach was implemented to optimize the model hyperparameters. The per-
formance of the model was evaluated using RMSE, MAPE, and R. The experimental
results showed that the LSTM model with 10 neurons provided a superior fit and high
prediction accuracy, followed by GRU with 50 neurons and CNN with 100 neurons.
Bhandari et al. Financial Innovation (2024) 10:75 Page 21 of 24

The outcome was further validated by a statistical analysis of the performance met-
rics. The proposed model can be tailored to other broad-market ESG indices for
which the data show similar characteristics.
In the near future, we plan to develop hybrid predictive models by combining the
implemented models with other neural network architectures such as transformers.
Another potential direction is the amalgamation of classical and deep learning model
architectures to build a new predictive model. We also plan to implement a hybrid
optimization algorithm that trains model parameters by combining local and global
optimizers. Finally, the implementation of evolutionary algorithms to achieve state-
of-the-art performance is a topic for future research.

Abbreviations
ESG Environment, Social, and Governmental
GRI Global Reporting Initiative
SASB Sustainability Accounting Standards Board
UNPRI United Nations Principles for Responsible Investment
UNSDG United Nations Sustainable Development Goals
ETF Exchange Traded Fund
ML Machine Learning
LSTM Long-Short Term Memory
GRU​ Gated Recurrent Unit
CNN Convolution Neural Networks
RMSE Root Mean Square Error
MAPE Mean Absolute Percentage Error
R Correlation Coefficient
MVP-ESG Modified Mean-Variance ESG
SRI Socially Responsible Investment
CSR Corporate Social Responsibility
ROE Return of Equity
ROA Return of Assets
VADER Valence Aware Dictionary and Sentiment Reasoner
DSSRI-I Double-Screening Socially Responsible Investment-I and
DSSRI-II Double-Screening Socially Responsible Investment-II
ELM Extreme Learning Machine
MSPR Maximum Sharpe Ratio
GA Genetic Algorithm
MSCI Morgan Stanley Capital International
PLS Partial Least Square
VIX CBOE Volatility Index
EFFR Effective Federal Funds Rate
UNRATE Civilian Unemployment Rate
UMCSENT Consumer Sentiment Index
USDX US Dollar Index
MACD Moving Average Convergence Divergence
RSI Relative Strength Index
SR Sharpe Ratio
SNPE Ticker of Xtrackers S &P 500 ESG ETF
EFIV Ticker of SPDR S &P 500 ESG ETF
ERTH Ticker of Invesco MSCI Sustainable Future ETF
SDG Ticker of iShares MSCI Global Sustainable Development Goals ETF
FNIDX Ticker of Fidelity International Sustainability Index Fund
VFTAX Ticker of Vanguard FTSE Social Index Fund
GPU Graphic Processing Unit

Acknowledgements
The authors would like to thank the Association of Nepalese Mathematicians in America for creating a collaborative
research opportunity that resulted in this work. We are also thankful to Google LLC for providing GPU supported open-
source cloud computing platform, Google Colab.

Author Contributions
All authors contributed to the study conception and design. Data collection and software development were performed
by HNB. Methodology and analysis were done by all authors together. The outline was created by HNB and all authors
contributed to the sections of the manuscript. All authors read and approved the final manuscript.
Bhandari et al. Financial Innovation (2024) 10:75 Page 22 of 24

Funding
No funding was received to assist with the preparation of this manuscript.

Availibility of data and materials

The data used in this work are open source, the readers can access the data following the manuscript. Codes will be
made available to public once the manuscript is accepted for publication.

Declarations
Competing interests
The authors have no competing interests to declare that are relevant to the content of this article.

Received: 14 October 2022 Accepted: 29 December 2023

References
Ahangar RG, Yahyazadehfar M, Pournaghshband H (2010) The comparison of methods artificial neural network with
linear regression using specific variables for prediction stock price in Tehran stock exchange. arXiv:​1003.​1457
Akhtaruzzaman M, Boubaker S, Umar Z (2022) COVID-19 media coverage and ESG leader indices. Finance Res Lett
45:102170
Albuquerque R, Koskinen Y, Zhang C (2019) Corporate social responsibility and firm risk: theory and empirical evi-
dence. Manage Sci 65(10):4451–4469
Anghel GDI (2015) Stock market efficiency and the MACD. Evidence from countries around the world. Procedia Econ
Finance 32:1414–1431
Baker M, Wurgler J (2007) Investor sentiment in the stock market. J Econ Perspect 21(2):129–152
Berg F, Lo AW, Rigobon R, Singh M, Zhang R (2023) Quantifying the returns of ESG investing: an empirical analysis
with six ESG metrics. Available at SSRN
Bernanke BS, Kuttner KN (2005) What explains the stock market’s reaction to Federal Reserve policy? J Finance
60(3):1221–1257
Bhandari HN, Rimal B, Pokhrel NR, Rimal R, Dahal KR (2022) LSTM-SDM: an integrated framework of LSTM implemen-
tation for sequential data modeling. Softw Impacts 14:100396. https://​doi.​org/​10.​1016/j.​simpa.​2022.​100396
Bhandari HN, Rimal B, Pokhrel NR, Rimal R, Dahal KR, Khatri RK (2022) Predicting stock market index using LSTM. Mach
Learn Appl 100320
Bock J (2018) Quantifying macroeconomic expectations in stock markets using google trends. arXiv:​1805.​00268
Brownlee J (2018) Better deep learning: train faster, reduce overfitting, and make better predictions. Ebook: Machine
Learning Mastery
Brownlee J (2018) Deep learning for time series forecasting: predict the future with MLPs, CNNs and LSTMs in python.
Ebook: Machine Learning Mastery
Brownlee J (2018c) How to develop convolutional neural network models for time series forecasting. Machine Learn-
ing Mastery
Chandra A, Thenmozhi M (2015) On asymmetric relationship of India volatility index (India VIX) with stock market
return and risk management. Decision 42(1):33–55
Chatzis SP, Siakoulis V, Petropoulos A, Stavroulakis E, Vlachogiannakis N (2018) Forecasting stock market crisis events
using deep and statistical machine learning techniques. Expert Syst. Appl. 112:353–371. https://​doi.​org/​10.​
1016/j.​eswa.​2018.​06.​032
Chen Q, Liu XY (2020). Quantifying ESG alpha using scholar big data: an automated machine learning approach. In:
Proceedings of the first ACM international conference on AI in finance, pp 1–8
Cho P, Lee M (2022) Forecasting the volatility of the stock index with deep learning using asymmetric Hurst expo-
nents. Fractal Fract 6(7):394
Chollet F (2017) Deep learning with python. Simon and Schuster, New York
Chong TT-L, Ng W-K (2008) Technical analysis and the London stock exchange: testing the MACD and RSI rules using
the FT30. Appl Econ Lett 15(14):1111–1114
Chong TT-L, Ng W-K, Liew VK-S (2014) Revisiting the performance of MACD and RSI oscillators. J Risk Financ Manag
7(1):1–12
Clementino E, Perkins R (2021) How do companies respond to environmental, social and governance (ESG) ratings?
Evidence from Italy. J Bus Ethics 171(2):379–397
D’agostino
√ R, Pearson ES (1973). Tests for departure from normality. Empirical results for the distributions of b 2 and
b . Biometrika 60(3):613–622
Daniali SM, Barykin SE, Kapustina IV, Mohammadbeigi Khortabi F, Sergeev SM, Kalinina OV, Senjyu T (2021) Predicting
volatility index according to technical index and economic indicators on the basis of deep learning algorithm.
Sustainability 13(24):14011
De Lucia C, Pazienza P, Bartlett M (2020) Does good ESG lead to better financial performances by firms? machine
learning and logistic regression models of public enterprises in Europe. Sustainability 12(13):5317
Engelhardt N, Ekkenga J, Posch P (2021) ESG ratings and stock performance during the COVID-19 crisis. Sustainability
13(13):7133
Eric D, Andjelic G, Redzepagic S (2009) Application of MACD and RVI indicators as functions of investment strategy
optimization on the financial market. Zbornik radova Ekonomskog fakulteta u Rijeci:casopis za ekonomsku
teoriju i praksu 27(1):171–196
Bhandari et al. Financial Innovation (2024) 10:75 Page 23 of 24

Farsio F, Fazel S (2013) The stock market/unemployment relationship in USA, China and Japan. Int J Econ Financ
5(3):24–29
Gary SN (2019) Best interests in the long term: fiduciary duties and ESG integration. U. Colo. L. Rev. 90:731
Gáeron A (2019) Hands-on machine learning with Scikit-Learn, Keras, and Tensorflow: concepts, tools, and tech-
niques to build intelligent systems. O’Reilly Media, California
Gers FA, Schmidhuber J, Cummins F (2000) Learning to forget: continual prediction with LSTM. Neural Comput
12(10):2451–2471
Gers FA, Schraudolph NN, Schmidhuber J (2003) Learning precise timing with LSTM recurrent networks. J Mach Learn
Res 3:115–143. https://​doi.​org/​10.​1162/​15324​43037​68966​139
Goyal M, Aggarwal K (2014) ESG index is good for socially responsible investor in India. Asian J Multidiscip Stud
2(11):92–96
Greff K, Srivastava RK, Koutník J, Steunebrink BR, Schmidhuber J (2017) LSTM: a search space odyssey. IEEE Trans
Neural Netw Learn Syst 28(10):2222–2232. https://​doi.​org/​10.​1109/​TNNLS.​2016.​25829​24
Guo T, Jamet N, Betrix V, Piquet L-A, Hauptmann E (2020) Esg2risk: A deep learning framework from ESG news to
stock volatility prediction. arXiv:​2005.​02527
Hochreiter S, Schmidhuber J (1997) Long short-term memory. Neural Comput 9(8):1735–1780
Indices SDJ (2016) S &P Dow Jones indices. Retrieved 12 April 2016
Jabeur SB, Khalfaoui R, Arfi WB (2021) The effect of green energy, global environmental indexes, and stock markets in
predicting oil price crashes: evidence from explainable machine learning. J Environ Manag 298:113511
Koo E, Kim G (2023) A new neural network approach for predicting the volatility of stock market. Comput Econ
61(4):1665–1679
Kou G, Chao X, Peng Y, Alsaadi FE, Herrera-Viedma E (2019) Machine learning methods for systemic risk analysis in
financial sectors. Technol Econ Dev Econ 25(5):716–742
Kou G, Olgu Akdeniz Ö, DinÇcer H, Yüksel S (2021) Fintech investments in European banks: a hybrid IT2 fuzzy multidi-
mensional decision-making approach. Financ Innov 7(1):39
Lee O, Joo H, Choi H, Cheon M (2022) Proposing an integrated approach to analyzing ESG data via machine learning
and deep learning algorithms. Sustainability 14(14):8745
Lee TK, Cho JH, Kwon DS, Sohn SY (2019) Global stock market investment strategies based on financial network
indicators using machine learning techniques. Expert Syst Appl 117:228–242
Lei J, Liu C, Jiang D (2019) Fault diagnosis of wind turbine based on long short-term memory networks. Renew
Energy 133:422–432. https://​doi.​org/​10.​1016/j.​renene.​2018.​10.​031
Li T, Kou G, Peng Y, Philip SY (2021) An integrated cluster detection, optimization, and interpretation approach for
financial data. IEEE Trans Cybern 52(12):13848–13861
Lin S-L, Jin X (2023) Does ESG predict systemic banking crises? A computational economics model of early warning
systems with interpretable multi-variable LSTM based on mixture attention. Mathematics 11(2):410
Lu X, Ma F, Wang J, Dong D (2022) Singlehanded or joint race? Stock market volatility prediction. Int Rev Econ
Finance 80:734–754
Margot V, Geissler C, de Franco C, Monnier B, Advestis F, Ossiam F (2021) ESG investments: filtering versus machine
learning approaches. Appl Econ Finance 8(2):1–16
Mittnik S, Robinzonov N, Spindler M (2015) Stock market volatility: identifying major drivers and the nature of their
impact. J Bank Finance 58:1–14
Murphy JJ (1999) Technical analysis of the financial markets: a comprehensive guide to trading methods and applica-
tions. Penguin, New York
Nabipour M, Nayyeri P, Jabani H, Shahab S, Mosavi A (2020) Predicting stock market trends using machine learn-
ing and deep learning algorithms via continuous and binary data; a comparative analysis. IEEE Access
8:150199–150212
Pokhrel NR, Dahal KR, Rimal R, Bhandari HN, Khatri RK, Rimal B, Hahn WE (2022) Predicting nepse index price using
deep learning models. Mach Learn Appl 100385
Qiu J, Wang B, Zhou C (2020) Forecasting stock prices with long-short term memory neural network based on atten-
tion mechanism. PLOS ONE 15(1):1–15. https://​doi.​org/​10.​1371/​journ​al.​pone.​02272​22
Raman N, Bang G, Nourbakhsh A (2020) Mapping ESG trends by distant supervision of neural language models.
Mach Learn Knowl Extract 2(4):453–468
Rimal B (2022) Financial time-series analysis with deep neural networks (Unpublished doctoral dissertation). Florida
Atlantic University
Rodríguez-González A, García-Crespo Á, Colomo-Palacios R, Iglesias FG, Gómez-Berbís JM (2011) Cast: Using neural
networks to improve trading systems based on technical analysis by means of the RSI financial indicator. Expert
Syst Appl 38(9):11489–11500
Ruan L (2018) Research on sustainable development of the stock market based on VIX index. Sustainability
10(11):4113
Sabbaghi O (2020) The impact of news on the volatility of ESG firms. Glob Finance J 100570
Schmidt AB (2022) Optimal ESG portfolios: an example for the Dow Jones index. J Sustain Finance Invest
12(2):529–535
Sen J, Chaudhuri TD (2018) Stock price prediction using machine learning and deep learning frameworks. In: Pro-
ceedings of the 6th international conference on business analytics and intelligence, Bangalore, India, pp 20–22
Serfling MA, Miljkovic D (2011) Time series analysis of the relationships among (macro) economic variables, the
dividend yield and the price level of the S &P 500 index. Appl Financ Econ 21(15):1117–1134
Tien NH, Jose RJS, Ullah SE, Thang HV (2021) The impact of world market on Ho Chi Minh city stock exchange in
context of COVID-19 pandemic. Turk J Comput Math Educ (TURCOMAT) 12(14):4252–4264
Tucker JJ III, Jones S (2020) Environmental, social, and governance investing: investor demand, the great wealth
transfer, and strategies for ESG investing. J Financ Serv Prof 74(3):56
Bhandari et al. Financial Innovation (2024) 10:75 Page 24 of 24

Umar Z, Abrar A, Zaremba A, Teplova T, Vo XV (2022) Network connectedness of environmental attention green and
dirty assets. Finance Res Lett 50:103209
Umar Z, Gubareva M (2021) The relationship between the COVID-19 media coverage and the environmental, social
and governance leaders equity volatility: a time–frequency wavelet analysis. Appl Econ 53(27):3193–3206
Vo NN, He X, Liu S, Xu G (2019) Deep learning for decision making and the optimization of socially responsible
investments and portfolio. Decis Support Syst 124:113097
Wang J, Kim J (2018) Predicting stock price trend using MACD optimized by historical volatility. Math Probl Eng
2018:1–12
Wang Z, Hong T, Piette MA (2020) Building thermal load prediction through shallow machine learning and deep
learning. Appl Energy 263:114683
Wilder JW (1978) New concepts in technical trading systems. Trend Research, New York
Winegarden W (2019) Environmental, social, and governance (ESG) investing: an evaluation of the evidence. Pacific
Research Institute
Xidonas P, Essner E (2022) On ESG portfolio construction: a multi-objective optimization approach. Comput Econ
1:25
Yu, G., Liu, Y., Cheng, W., Lee, C.-T. (2022). Data analysis of ESG stocks in the Chinese stock market based on machine
learning. In: 2022 2nd international conference on consumer electronics and computer engineering (ICCECE), pp
486–493
Zhang J, Chen X (2021) Socially responsible investment portfolio construction with a double-screening mechanism
considering machine learning prediction. Discrete Dyn Nat Soc 2021

Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.