Time Series Analysis
PROF. DR. MUHAMMAD AZAM
Outline
Time Series
Analysis of Time Series
Historigram
Components of Time Series
Secular Trend
Cyclical Variations
Seasonal Variations
Irregular Variations
Measurement of Trend
Method of Free-Hand Curve
Method of Semi Averages
Method of Moving Averages
Method of Least Squares
Time Series Day Cases Pakistan
1-Apr 2118
A time series is a set of observations 2-Apr 2421
recorded according to some period of
time. The observations are usually 3-Apr 2686
recorded at equal intervals of time. Time 4-Apr 2818
periods may be days, months, quarters,
years or decades etc. 5-Apr 3157
For example, record of total number of 6-Apr 3766
COVID-19 patients in Pakistan on daily 7-Apr 4035
basis. This is called Daily Time Series.
8-Apr 4263
Population of a country may be
counted after every 10 years. So it will be 9-Apr 4489
a decennial time series. 10-Apr 4695
Time Series
Time Values (𝑦𝑡 )
Mathematically a time series is defined
𝑡1 𝑦1
by the values 𝑦1 , 𝑦2 , 𝑦3 … , 𝑦𝑛 at time 𝑡1 , 𝑡2 , 𝑡2 𝑦2
𝑡3 … , 𝑡𝑛 . So 𝑡3 𝑦3
. .
𝑦𝑡 = 𝑓 𝑡 + 𝑈𝑡
. .
. .
𝑡𝑛 𝑦𝑛
Analysis of Time Series
The forces acting on time series are called variations or fluctuations. The time series
variable is made up of other variables which are 𝑇, 𝑆, 𝐶 𝑎𝑛𝑑 𝐼. Where 𝑇 stands for Trend, 𝑆
for Seasonal, 𝐶 for Cyclical and 𝐼 for Irregular. So 𝑦𝑡 can be written as
𝑦𝑡 = 𝑇𝑆𝐶𝐼
Which is called multiplicative model of time series. The additive model of time series can
be written as
𝑦𝑡 = 𝑇 + 𝑆 + 𝐶 + 𝐼
The measurement of Trend, Seasonal, Cyclical or Irregular variations is called Analysis of
Time Series
Historigram
Graph of the time series is called Time Series Plot of Cases Pakistan
14000
historigram. Suppose 𝑡1 , 𝑡2 , 𝑡3 … , 𝑡𝑛 be
12000
time periods with corresponding values 10000
Cases Pakistan
of variable Y are 𝑦1 , 𝑦2 , 𝑦3 … , 𝑦𝑛 . Then the 8000
time periods are taken on X-axis and Y-
6000
4000
values on Y-axis. All the values are 2000
plotted against their respective time. 0
Day 21 1 11 21 1 11 21
The plotted points are joined together Month Feb Mar Apr
by straight lines to get the graph of the
historigram.
Components of Time Series
There are four types of variations in time series
which are given below:
1. Secular Trend: It means a long period trend. Such
type of trend shows the movement of time series in
one direction over a long period (several years or
decades). Secular trend shows the general pattern
of increase or decease in an economic activity.
These movements are free of sudden jumps or
swings. These movements may be slow or fast but
they are regular in nature.
Components of Time Series
2. Cyclical Variations: These are long-term swings
(ups and downs) which takes the upward or
downward form. These movements are business
cycles which pass through the stages of boom
(prosperity), recession (decline), depression and
recovery. The distance between two booms is called
cycle and the time involved between two booms is
called span of the cycle.
Components of Time Series
Economic Boom: An economic boom refers to a
period of increased commercial activity within
either a business, market, industry, or economy as
a whole. For an individual company,
a boom means rapid and significant sales growth,
while a economic boom for a country is marked
by significant GDP growth.
Components of Time Series
3. Seasonal Variations: These are short-term variations but they appear regularly in the nature.
By “season” we meant a specific period of the year. These specific periods of the year may be
“A Holiday”, “Ramadan”, “Week Ends”, “Eid Days”, “Summer”, “Winter” etc. Therefore the
elements of increase or decease in time series which is totally due to changes in seasons is
called Seasonal Variations. The increase in prices of fruits in Ramadan, increase in demand of
Ice-Cream in summer etc. are the examples of Seasonal Variations.
4. Irregular Variations: Such type of variations appears all of a sudden in time series. Such as
wars, floods, earthquakes, strikes, epidemics etc. Rise in prices during wars, delay in production
due to labor strikes etc. are the examples of irregular variations. These variations can’t be
controlled. They are also called random or accidental variations.
