Scribd
Upload
54 views24 pages

Unit 8 Time Series

This document discusses various forecasting techniques including qualitative methods like executive opinion and quantitative time series analysis methods. It covers components of a time series like trend, seasonality, cycles and irregular variations. Specific quantitative time series forecasting techniques covered include freehand method, smoothing techniques like semi-averages and moving averages, least squares method and exponential smoothing. It also discusses weighted moving averages and calculating trend values using different weighted moving averages. Finally, it outlines some utilities of time series analysis like understanding past behavior and facilitating planning and comparison.

Uploaded by

Shrawani Deshmukh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
54 views24 pages

Unit 8 Time Series

This document discusses various forecasting techniques including qualitative methods like executive opinion and quantitative time series analysis methods. It covers components of a time series like trend, seasonality, cycles and irregular variations. Specific quantitative time series forecasting techniques covered include freehand method, smoothing techniques like semi-averages and moving averages, least squares method and exponential smoothing. It also discusses weighted moving averages and calculating trend values using different weighted moving averages. Finally, it outlines some utilities of time series analysis like understanding past behavior and facilitating planning and comparison.

Uploaded by

Shrawani Deshmukh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 24

FORECASTING

TECHNIQUES
Methods of Forecasting
Qualitative methods
◦ 1. Executive opinion
◦ 2. Panel Judgement
◦ 3. Delphi
◦ 4. Marketing research
◦ 5. Past analogy

Quantitative methods
1. Time series analysis
Freehand
Smoothening technique
Exponential smoothening
2. Causal analysis
Regression trend analysis
COMPONENTS OF A TIME SERIES.

Secular Trend T
Seasonal Variations S
Cyclical Variations C
Irregular Variations I
Quantitative methods of time series forecasting

Freehand or Graphical method


Smoothening technique:
◦ Semi-average method
◦ Moving average method

Least squares method.


Freehand or Graphical method
1. Plot the time series on a graph paper.
2. Examine carefully the direction of dots.
3. Draw a straight line according to personal judgement.
Fit a trend line to the following data using the Freehand
method and predict values for 1998 & 1999.

SUGAR PRODUCTION
YEAR
(Million Tonnes)

1992 10
1993 35
1994 30
1995 55
1996 45
1997 60
90

80

70

60

50

40

30

20

10

92 93 94 95 96 97 98 99
Fit a trend line to the following data using the Freehand
method and predict values for 2009 & 2010.

YEAR Sales
(Millions)
2001 5
2002 15
2003 10
2004 25
2005 30
2006 20
2007 35
2008 45
Smoothing Techniques: The main objective of the smoothing
technique is to smooth out the random variations due to the
irregular fluctuations in time series data.

Method of SEMI AVERAGES


1. Divide the data in two equal parts. In case of odd years, omit the middle year.
2. Obtain the average of each part.
3. Plot the two points against the midpoint of class interval on a graph.
4. Joint the two points to get a trend line.
Fit a trend line to the following data using the semi
averages method and predict values for 2009 & 2010.

Sales
YEAR (Millions)

1993 102
1994 105
1995 114
1996 110
1997 108
1998 116
1999 112
Method of MOVING AVERAGES

There can be two ways to calculate moving averages.


1. 3 year, 5 year or 7 year moving averages. These are called odd
year moving averages.
Or
2. 2 year, 4 year, 6 year or 8 year moving averages.

There is a slight difference in these two ways.


Calculate the 3 year moving averages of the production
figures given below.

YEAR PRODUCTION YEAR PRODUCTION


1985 15 1993 63
1986 21 1994 70
1987 30 1995 74
1988 36 1996 82
1989 42 1997 90
1990 46 1998 95
1991 50 1999 102
1992 56
Construct 5 year moving averages of the number of
students studying in a college.

YEAR No. of students YEAR No. of students

1990 332 1995 405


1991 317 1996 410
1992 357 1997 427
1993 392 1998 405
1994 402 1999 438
Calculate the trend values by taking 4 year moving
averages.

YEAR VALUE YEAR VALUE

1984 12 1991 100

1985 25 1992 82

1986 39 1993 65

1987 54 1994 49

1988 70 1995 34

1989 87 1996 20

1990 105 1997 7


Calculate the trend values by taking 4 year moving
averages.

YEAR VALUE YEAR VALUE

1984 50.0 1989 38.1

1985 36.5 1990 32.6

1986 43.0 1991 41.7

1987 44.5 1992 41.1

1988 38.9 1993 33.8


WEIGHTED MOVING AVERAGES

Generally weighted moving average is used to forecast


trend figures.
WMA gives higher weightage to recent figures.
Calculate the trend values using 3 year WMA for the
following data. Weights are to be assigned in order 1, 2,
3.

YEAR SALES YEAR SALES


2001 10 2008 18
2002 12 2009 20
2003 12 2010 18
2004 14 2011 24
2005 16 2012 28
2006 18
2007 22
Year Sales WT Wtd Sales 3 Y WMT 3 Y WMA
01 10 1 10 ---- ----
02 12 2 24 70 11.66
03 12 3 36 74 12.33
04 14 1 14 82 13.66
05 16 2 32 100 16.66
06 18 3 54 108 18
07 22 1 22 112 18.66
08 18 2 36 118 19.66
09 20 3 60 114 19
10 18 1 18 126 21
11 24 2 48 150 25
12 28 3 84 ---- ----
Calculate the trend values using 5 year WMA for the
following data. Weights are to be assigned in order 1, 2,
2, 3, 3.

YEAR SALES YEAR SALES


1990 18 1997 32
1991 20 1998 28
1992 21 1999 36
1993 26 2000 34
1994 22 2001 35
1995 24 2002 44
1996 30 2003 46
2004 42
For the following data, verify that the 5 yearly weighted
moving average with weights 1,2,3,3,1 resp. is equivalent to 4
yearly centered moving average.

Year 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

Sales 5 3 7 6 4 8 9 10 8 9 9
LEAST SQUARES METHOD
EQUATION OF SRTAIGHT TREND LINE
Y = a + bX
Normal Equations for obtaining the values of a and b
are as follows

(i) ∑ Y = Na + b ∑ X

(ii) ∑ XY = a ∑X + b ∑ X²

N = Number of years,
X = Converted value for years.
Exponential Smoothing method
Exponential smoothing is a time series forecasting method for
univariate data.
Exponential moving average gives a higher weighting to recent prices,
while the simple moving average assigns equal weighting to all values.
Exponential smoothing forecasting methods are similar in that a
prediction is a weighted sum of past observations, but the model
explicitly uses an exponentially decreasing weight for past observations.
UTILITY OF TIME SERIES ANALYSIS.

Helps in understanding past behaviour.


Helps in planning future operations.
Helps in evaluating current accomplishments.
Facilitates comparison.

You might also like