Time Series Analysis by Salah Uddin

COURSE CODE: EMIS-506, BUSINESS STATISTICS
DR. K. M. SALAH UDDIN, ASSOCIATE PROFESSOR

DEPARTMENT OF MANAGEMENT INFORMATION SYSTEMS (MIS)
UNIVERSITY OF DHAKA
TIME SERIES ANALYSIS
Time Series Analysis

A time series is a set of observed values, such as production or sales
data, for a sequentially ordered series of time periods. Examples of such
data are sales of a particular product for a series of months and the
number of workers employed in a particular industry for a series of
years.
Mathematically,
A time series is a set of observations taken at specified times, usually at
equal intervals. Mathematically, a time series is defined by the values
Y1,,Y2,..of a variable Y at times t1, t2,.. Thus Y is a function of t ,
symbolized Y=F(t).
Role of Time Series Analysis
Time series analysis is of great significance in decision-making for the
following reasons:
1. It helps in the understanding of past behavior by observing data
over a period of time; one can easily understand what changes

UNIVERSITY OF DHAKA
have taken place in the past. Such analysis will be extremely

helpful in predicting future behavior.
2. It helps in planning future operations. In fact, the greatest
potential of time series lies in predicting an unknown value of the
series. From this information intelligent choices can be made
concerning capital investment decisions, decisions concerning
production and inventory etc.
3. It helps in evaluating current accomplishments. The actual
performance can be compared with the expected performance and
the cause of variation analyzed. For example, if expected sales for
2013 were 20 lacs colored T. V. sets and the actual sales were only
19 lacs; one can investigate the cause for shortfall in achievement.
4. It facilitates comparison. Different time series are often compared
and important conclusions drawn therefrom.
Method of Least Squares
The method of least squares is the most frequent basis used for
identifying the trend component of the time series by determining the
equation for the best-fitting trend line. When this method is applied, a
trend line is fitted to the data in such a manner that the following two
conditions are satisfied:

UNIVERSITY OF DHAKA
1) (Y-Yc)=0
i.e., the sum of deviation of the actual values of Y and the computed
values of Y is zero.
2) (Y-Yc)2 is least,
i.e., the sum of the squares of the deviations of the actual and computed
values is least from this line. That is why this method is called the
method of least squares.
The straight line trend is represented by the equation
Yc=a+b X
Where Yc denotes the trend (computed) values to distinguish them from
the actual Y values, a is the Y intercept or the value of the Y variable
when X=0, b represents slope of the line or the amount of change in Y
variable that is associated with a change of one unit in X variable. The X
variable in time series represents time.
In order to determine the value of the constants a and b, the following
two normal equations are to be solved:
Y=N a+b X
XY= aX+bX2

UNIVERSITY OF DHAKA
The time variable is measured as a deviation from its mean. Since

X=0, the above two normal equations would take the form:
Y=N a
XY= bX2
the values of a and b can now be determined easily.
Since
Y=N a
a
Since
Y
Y
N
XY= bX2
XY
2
X
The constant a give the arithmetic mean of Y and the constant b

indicates the rate of change.
Problem 01: Fit a straight line trend by the method of least squares to the
following data and find the trend values:
Years
Sale of air conditioners (in
lakhs )
2003
2004
2005
2006
2007
2008
10
13
16
21
24
30

UNIVERSITY OF DHAKA
Solution: Fitting straight line trend by the method of least squares

Year
X2
Sales (in
Deviation
lakhs )
from 2005
2003
10
-2
-20
9.143
2004
13
-1
-13
13.086
2005
16
16
17.029
2006
21
+1
21
20.972
2007
24
+2
48
24.915
2008
30
+3
90
28.858
X=3
XY=126
N=6
Y=114
XY
Trend values
Yc
X2=19 Yc=114.003
The equation of the straight line trend is : Yc=a+b X

Since X is not zero, we have to solve the two normal equations
simultaneously,
Y=N a+b X
XY= aX+bX2
114=6a+3b
(i)
126=3a+19 b
(ii)
Multiplying equation (ii) by 2

UNIVERSITY OF DHAKA
114=6a+3b
252=6a+38 b
-35 b=-138
Or, b=3.943
Putting the value of b in equation (i)
114=6a+33.943
6a=114-11.829
6a=102.171
a=17.029
the equation of the straight line trend is :
Y=17.029+3.943 X
Y2003=17.029+3.943 (-2)=9.143
Y2004=9.143+3.943=13.086
Second Degree Parabola:
The simplest example of the non-linear trend is the second degree
parabola, the equation of which is written in the form:

UNIVERSITY OF DHAKA
Yc=a+b X+c X2
when the numerical values for constants a, b and c have been derived,
the trend value for any year may computed substituting the value of X
for that year. The values of a, b and c can be obtained by solving the
following three normal equations simultaneously:
i.
Y=N a+b X+cX2
ii.
XY= aX+bX2+cX3
iii.
X2Y= aX2+bX3+cX4
When time origin is taken between two middle years X and X3 would
be zero. In that case the above equations are reduced to
i.
Y=N a +cX2
ii.
XY= bX2
iii.
X2Y= aX2 +cX4
Problem 02: The price (in dollars) of a commodity during 2003-2008 is

given below. Fit a parabola Y=a+bX+cX2 to this data. Estimate the
price of commodity for the year 2007 and also compute the trend values.
Year
2003
`2004
2005
2006
2007
2008
Price
100
107
128
140
181
192

UNIVERSITY OF DHAKA
Solution: To determine the value of a,b and c, we solve the following

normal equations:
Y=N a+b X+cX2
XY= aX+bX2+cX3
X2Y= aX2+bX3+cX4
Year
Yc
Price
X2
X3
X4
XY
X2Y
Trend
values
2003
100
-2
-8
16
-200
400
97.744
2004
107
-1
-1
-107
107
110.426
2005
128
126.680
2006
140
+1
140
140
146.506
2007
181
+2
16
363
724
169.904
2008
192
+3
27
81
576
1728
196.874
N=6
Y=8
X=3
X2
X3
X2Y=3099
Yc=848.134
=19
=27
48
X4=115 XY=77
1
Substituting these values

848=6a+3b+19c
(i)
771=3a+19b+27c
(ii)
3099=19a+27b+115c
(iii)

UNIVERSITY OF DHAKA
Solving equations (i) and (ii), we get

35b+35c=694
(iv)
Multiplying equation (ii) by 19 and Equation (iii) by 3 and subtracting,

we get
53.52=280b+168c
(v)
Solving equations (iv) and (v), we get

c=1.786
substituting the value of c in equation (iv), we get
b=18.04
putting the value of b and c in equation (i), we get
a=126.68
Thus, a=126.68, b=18.04 and c=1.786
Substituing the values in the equation
Y=126.68+18.04X+1.786 X2
When X=2, Y2007=126.68+18.04(2)+1.786(2)3
=169.904

UNIVERSITY OF DHAKA
Exponential or Logarithmic Trend:

The equation of the exponential curve is of the form
Y=a bx
Putting the equation logarithmic form, we get
log Y=log a+ X log b
when plotted on a semi-logarithmic graph, the curve gives a straight line.
However, on an arithmetic chart the curve gives a nonlinear trend. In
order to find out the values of a and b, the two normal equations to be
solved are:
log Y=N log a +log b X
(X. log Y)=log a X +log b.X2
when deviations are taken from middle year, i. e., X=0, the above
equations take the following form:
log Y=N log a
(X. log Y)= log b.X2

UNIVERSITY OF DHAKA
log a
log Y
X . log Y
; and log b
2
N
X
Take the antilogs of these expressions to arrive at the actual trend values.
Problem 04:
Fit a logarithmic straight line to the following data:
Years
2003
2004
2005
2006
2007
2008
64
70
75
82
88
95
Production (m. tons of steel)
Solution:
Year
Production Deviation from

Y
log Y
X2
X. log Y
2006
X
2003
64
-3
1.8062
-5.4186
2004
70
-2
1.8451
3.6902
2005
75
-1
1.8751
-1.8751
2006
82
1.9138
2007
88
+1
1.9445
1.9445
2008
95
+2
1.9777
3.9554
log Y=11.3624
X2=19
X. log
N=6
Y=474
X=-3
Y=-5.084
The logarithmic straight line trend is given by

log Y=log a+ X log b

UNIVERSITY OF DHAKA
The two normal equations are:

log Y=N log a +log b X
(X. log Y)=log a X +log b.X2
Substituting the values
11.3624=6 log a-3 log b
(i)
-5.084=-3 log a+19 log b
(ii)
Multiplying equation (ii) by 2 and adding to (i)

11.3624=6 log a-3 log b
-10.168=-6 log a+38 log b
35 log b=1.1944
log b=1.1944/35=0.034
Putting the value of log b in equation (i)
11.3624=6 log a-30.034
=> 6 log a=11.4644
=> log a=1.911
Hence, log Y=1.911-0.034 X

Time Series Analysis by Salah Uddin

Transféré par

Informations du document

Description originale:

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Time Series Analysis by Salah Uddin

Transféré par

Droits d'auteur :

Formats disponibles

COURSE CODE: EMIS-506, BUSINESS STATISTICS

DR. K. M. SALAH UDDIN, ASSOCIATE PROFESSOR