STATISTICAL TECHNIQUE FOR MANAGERS
MODULE 4: TIME SERIES ANALYSIS AND FORECASTING 10 hrs
Time Series Analysis: Components, Models of Time Series–Additive, Multiplicative
and Mixed models- Trend analysis: Free hand curve, Semi averages, Moving
averages, Least Square method. Model Assumptions and Diagnostics, Practical
Business Forecasting Techniques. Data Analysis and Decision Making, Exploratory
Data Analysis (EDA).
Time Series Analysis:
Time Series: A sequence of data points,
observations or measurements taken at
successive, equally spaced time intervals.
Time series analysis is the study of data
collected or recorded over successive time
intervals to identify patterns, trends and behaviors
that change over time.
Time Series Analysis:
1. Components of Time Series: Trend, Seasonality, Cyclical,
Irregular/Random.
2. Models of Time Series:
a) Additive Model: The components are added together
linearly.
b) Multiplicative Model: The components are multiplied
together.
c) Mixed Models: A combination of both additive and
multiplicative models.
Trend
[Link] trend component represents the long-term
movement or direction of the data over time.
[Link] shows whether the data is increasing,
decreasing or staying relatively constant over
time.
[Link] can be linear, where there constant rate
of change or non-linear and where the rate of
change varies.
[Link] are essential for understanding the overall
pattern of the data and making forecasts.
Seasonality
[Link] refers to patterns in the data that
repeat at regular intervals and typically within a
year.
[Link] patterns are often influenced by seasonal
factors such as weather, holidays or cultural
events etc…
[Link] and accounting for seasonality is
crucial for making accurate forecasts, as ignoring
it can lead to biased predictions.
Cyclical
1. Cyclical components represent longer-term
fluctuations in the data that are not of a fixed period
like seasonality.
2. These fluctuations often correspond to economic
cycles, business cycles or other periodic fluctuations
in the underlying data-generating process.
3. Unlike seasonality, cyclical patterns do not have a fixed
duration and can vary with duration over time.
4. Identifying cyclical patterns can be challenging as they
may overlap with trends or other components of the
time series.
Irregular/Random
1. The irregular component, also known as the residual or
noise, represents the random fluctuations or variations
in the data that cannot be attributed to the trend,
seasonality or cyclical patterns.
2. These fluctuations can arise due to random events,
measurement errors or other unpredictable factors.
3. The irregular component is often assumed to follow a
stationary stochastic process with zero mean and
constant variance.
4. While it's not possible to predict or model the irregular
component directly, understanding its characteristics is
important for assessing the overall uncertainty in the
data and the reliability of forecasts.
Models of Time Series
a) Additive Model: The components are added
together linearly.
b) Multiplicative Model: The components are
multiplied together.
c) Mixed Models: A combination of both
additive and multiplicative models.
• Y = T*S*C*I (multiplicative model)
• Y = T+S+C+I (additive model)
• Y = T+S*C*I (mixed model)
Methods of Fitting a Straight Line to a Time
Series
[Link] hand method
[Link] method of semi-averages
[Link] method of moving
averages
[Link] method of least squares.
Free Hand Method:
The method of semi-averages
Time (months) Sales (units) Semi-Average
1 100 -
2 110 (100+110)/2 = 105
3 120 (110+120)/2 = 115
4 115 (120+115)/2 = 117.5
5 130 (115+130)/2 = 122.5
Data of Daily Temperature Readings
Day Temperature (°C) Semi-Average
1 20 -
2 18 19
3 19 18.5
4 22 20.5
5 21 21.5
6 20 20.5
7 23 21.5
Calculate:
Method of Moving Averages
SN Date Gold Price (USD/ounce) 5-Day MA 10-Day MA
1 2024-03-01 1800 - -
2 2024-03-02 1820 - -
3 2024-03-03 1815 - -
4 2024-03-04 1832 - -
5 2024-03-05 1825 1818.4 -
6 2024-03-06 1840 1826.4 -
7 2024-03-07 1836 1829.6 -
8 2024-03-08 1822 1831 -
9 2024-03-09 1815 1827.6 -
10 2024-03-10 1808 1824.2 1821.3
11 2024-03-11 1795 1815.2 1819.8
12 2024-03-12 1802 1808.4 1817.4
13 2024-03-13 1810 1806 1816.5
14 2024-03-14 1820 1807.4 1817.3
15 2024-03-15 1835 1812.4 1819.4