3. Session Objectives
โข Time Series Analysis for Accurate Forecasting
โข Seasonality and Trend Analysis
โข Regression Analysis for Predictive Modeling
โข Combining Forecasting Methods for Robust Predictions
4. Think a minute
Letโs say you are running a business selling winter coats. Would you like
to diversify your business by starting an ice-cream manufacturing
business?
5. How are we going to predict the sales for next
January?
6. Time Series Analysis - Components
A data set of observations recorded in equal intervals over a period of
time (CIMA).
A time series is a sequence of data points that occur in successive order
over some period of time (Investopedia).
Key Terms:
Trend (T): The underlying general movement over time.
Seasonal Variation (SV): A regular variation within the data related to
calendar or season. Eg: Weather seasons.
Cyclical variation (C): A variation within the data related to a cyclical
pattern. Eg: Economic cycle.
Residual variation (R): Irregular or unpredictable changes in the data.
Eg: Impact of cyclone or war.
7. Models used in Time Series Analysis
Addictive Model: It assumes that all the components are independent
of each other.
TS = T+SV+C+R
Multiplicative Model: It assumes that the components are dependent
of each other.
TS = T*SV*C*R
8. Example - 01
Weeks Monday Tuesday Wednesday Thursday Friday
Week 01 160 174 164 180 150
Week 02 180 190 186 204 170
Week 03 210 224 206 232 190
Week 04 240 254 236 262 210
The total sales of Anbu plc for each day of a month is given below in โ$
millionsโ.
Identify the trend and calculate the average seasonal variation using
addictive model.
19. Example - 02
Weeks Monday Tuesday Wednesday Thursday Friday
Week 01 160 174 164 180 150
Week 02 180 190 186 204 170
Week 03 210 224 206 232 190
Week 04 240 254 236 262 210
The total sales of Anbu plc for each day of a month is given below in โ$
millionsโ.
Identify the trend and calculate the average seasonal variation using
multiplicative model.
31. Example 03
Years Quarter 01 Quarter 02 Quarter 03 Quarter 04
2020 160 174 164 180
2021 180 190 186 204
2022 210 224 206 232
The total sales of Anbu plc for each qua of 2020-2022 is given below in
โ$ millionsโ.
Calculate the sales for the first quarter of 2023 using Excel.
33. Activity 01
The ice cream sales (units in thousands) of Arasan limited for the years
2021 to 2023 are given below.
Estimate the sales for the seasons in 2024 using,
1. Addictive model
2. Multiplicative model
Summer Spring Autumn Winter
2021 500 450 300 120
2022 600 550 400 220
2023 700 650 500 320
34. Activity 02
The daily production level (units in thousands) of the employees of
Alfred manufacturers for the month of January is given below.
Estimate the sales for the days in February using multiplicative model
and regression analysis in Excel.
Monday Tuesday Wednesday Thursday Friday
Week 01 1200 1100 1130 1080 780
Week 02 1600 1500 1530 1480 1180
Week 03 2000 1900 1930 1880 1580
Week 04 1800 1700 1730 1680 1380
35. Activity 03
The total production cost and the total revenue details of Hassan
traders for the last 24 months is given in the next slide.
Calculate the total profit when the production level of,
A = 1300
B = 800
C = 1180
37. Summary
We learnt to,
โข Use the time series analysis to forecast the future estimates using addictive
and multiplicative models.
โข Understand the terms trend, seasonal variations, cyclical variation and
random variation.
โข Use regression analysis for predictive modeling.
โข combine forecasting methods for robust predictions.