TOPIC 3 – TIME SERIES ANALYSIS
Learning objective – WEEK 7
At the end of this topic student will able to:
1. Define the time series analysis
2. Identify the component of the time series
3. Use the moving average method for trend analysis (3MA, 5MA &4MA)
4. Calculate the seasonal variations
5. Compute the seasonal index and interpret its meanings
6. Forecast using a multiplicative model
A time series data is a set of numerical data for a certain variable that can describe the
upward and downward movements of the variable over time.
Time is taken as independent variable while other variables as dependent variable.
Time can be as yearly, monthly, weekly, quarterly or daily.
Observation of upward and downward patterns can be analyze using time series
analysis.
3.1 Component of time series.
Four components of a time series data are, trend, cyclical variations, seasonal
variations and irregular or random variation.
Trend (T)
It is the long term movement in a time series. It describes the pattern of the movement
of time series data either increases or decreases. Trend can be linear of non linear.
Cyclical variation (C)
It is long term movement in a time series data that repeat over time periods of differing
length usually longer than one year. Business cycles with their stages of prosperity,
depression, recession and recovery are important examples of cyclical variation.
Seasonal variation (S)
It is short term movement in a time series data. These movement repeat themselves
within a fixed period of one year. Many economic series have highs and lows due to
some special events occurring in one year.
Irregular or random variation (I)
It is movement that is unpredictable that place by chance or randomly. The destruction
of properties due to floods, fires, earthquakes etc are example of irregular variation.
3.2 Multiplicative time series model
The multiplicative model assumes that changes in variable Y in a series data are
caused by the product of the four components, namely Y=T x C x S X I. The trend in
units buts other three components in percentages.
Y = the time series variable
T = value of the trend component
C = value of the cyclical component
S = value of the seasonal component
I = value of the irregular or random component
3.3 Trend Analysis (Moving Average Method)
1. Estimate trend values using moving average method
- Odd MA – 3MA and 5MA
- Even MA – 4MA
2. Develop trend equation
TREND EQUATION (T) = a +bt
a = last trend value
last trend value-first trend value
b=
no of trend values-1
3. Computing seasonal indices
1. After obtaining the trend values using moving average method, calculate seasonal variation
using
actual data (Y)
Seasonal variation=
trend value (T)
2. Tabulate the seasonal variation according to time period (seasonal index table)
3. Find the average of each column. Find total mean of the variation.
4. Calculate correction factor (CF) using ( if 4 moving average)
4
CF=
total average of the variation
5. calculate the seasonal factor (SF) using
SF=each mean column × CF
6. calculate the Seasonal index (SI) using
SI=Seasonal factor ( SF ) ×100
4. Interpreting seasonal index value
If SI > 100% indicates higher (above) than trend values. If SI<100% indicates lower (below) than
trend values.
5. the dependent variable (Y)
Forecasting is the process of predicting the future using the data that are recorded in the past.
The forecasting will use multiplicative model.
Y=T ×S ×C ×I
SI
Y ≈ T ×S = (a +bt) ×
100
3.3.1& 3.3.2 Odds Moving Average (3MA, 5MA) & Even Moving Average (4MA)
Example: JUN2018
YEAR TERM ACTUAL 3 MOVING TREND 3 MOVING SEASOANAL
(Y) TOTAL AVERAGE (T) (Y/T)
2013 1 1.6
2 1.8
3 1.4
2014 1 1.5
2 1.5
3 1.7
2015 1 0.8
2 1.2
3 1.1
2016 1 1.3
2 2.4
3 1.6
Year Term 1 Term 2 Term 3
2013 -
2014
2015
2016 -
2017
2018 TOTAL
Total MEAN
Mean
Correction
factor
Seasonal
Factor
Seasonal
index
5 MOVING AVERAGE
YEA TERM ACTUAL 5 MOVING TREND 5 SEASOANAL
R (Y) TOTAL MOVING VARIATION =
AVERAGE (T) (Y/T)
2013 1 1.6
2 1.8
3 1.4
2014 1 1.5
2 1.5
3 1.7
2015 1 0.8
2 1.2
3 1.1
2016 1 1.3
2 2.4
3 1.6
Year Term 1 Term 2 Term 3 TOTAL
2013
2014
2015
2016
Total MEAN
Mean
CF
SF
SI
4 Moving Average
The sales of water filter (RM 000) at Rayyan Electrical Sdn Bhd from 2017 to 2019 is as follows:
Quarter
Year
1 2 3 4
2017 23 18 23 17 23+18+ 23+17
2018 18 22 19 15
2019 16 20 21 19
b) Compute the trend values using the moving-average method.
Total Centered
Actual data amount total Trend(T)=Centere Seasonal =
year quarter Y moved amount d 4 MA Y/T
2017 1 23
157
2 18 8
3 23
4 17
2018 1 18
2 22
3 19
4 15
2019 1 16
2 20
3 21
4 19
TREND EQUATION (T) = a +bt
a = last trend value = 18.5000
last trend value-first trend value 18.5000−19.6250
b= = =−0.1607
no of trend values-1 8−1
Therefore,
T = 18.5000 – 0.1607 t
Example 3 ( Cont )
b) calculate the seasonal indices
Total Centered Seasonal
Actual amount total Trend(T)=Centere variation=
year quarter data Y moved amount d 4 MA Y/T
2017 1 23
23
2 18 19 .6250
3 23
4 17
2018 1 18
2 22
3 19
4 15
2019 1 16
2 20
3 21
4 19
Seasonal table
Total
Quarter
Year mean
1 2 3 4
2017
2018
2019
2020
Total
Mean
CF
SF
SI
c) interpret meaning for seasonal indices for second and fourth quarter
S2= 112.53
12.53% higher than trend values
S4 = 85.70
14.3% lower than trend values.
d) forecast the total sales for the second quarter of 2020.
Second quarter of 2020 (t) is 4
Trend values = 18.5000 – 0.1607 (4) = 17.8572
Seasonal index for second quarter is 112.53
Therefore,
SI
Y=T ×S= ( a+bt ) ×
100
112.53
Y=17.8572 × =20.0927 ≈20.10
100
