Notes On Forecasting

28/01/16
Introduction
Chapter 5
n Managers are always trying to reduce
uncertainty and make better estimates of what

will happen in the future
n This is the main purpose of forecasting
n Some firms use subjective methods
n Seat-of-the pants methods, intuition,
experience
n There are also several quantitative techniques
n Moving averages, exponential smoothing,
trend projections, least squares regression
analysis`
Forecasting
To accompany
Quantitative Analysis for Management, Tenth Edition,
by Render, Stair, and Hanna
Power Point slides created by Jeff Heyl
2008 Prentice-Hall, Inc.

Introduction
Introduction
n Eight steps to forecasting :
n These steps are a systematic way of initiating,
1. Determine the use of the forecastwhat

objective are we trying to obtain?
2. Select the items or quantities that are to be
forecasted
3. Determine the time horizon of the forecast
4. Select the forecasting model or models
5. Gather the data needed to make the
forecast
6. Validate the forecasting model
7. Make the forecast
8. Implement the results
n
n
n
n
54
Time-Series Models
Forecasting
Techniques
Time-Series
Methods
designing, and implementing a forecasting

system
When used regularly over time, data is collected
routinely and calculations performed
automatically
There is seldom one superior forecasting system
Different organizations may use different
techniques
Whatever tool works best for a firm is the one
they should use
Assumptions
n The future event is determined by the past.
n Past data are available
53
Forecasting Models
Qualitative
Models
52
n Time-series models attempt to predict the
future based on the past
Causal
Methods
Delphi
Methods
Moving
Average
Regression
Analysis
Jury of Executive
Opinion
Exponential
Smoothing
Multiple
Regression
Sales Force
Composite
Trend
Projections
Consumer
Market Survey
Decomposition
n Common time-series models are

n Nave
n Simple moving average and weighted moving
average
n Exponential smoothing
n Trend projections
n Decomposition
n Regression analysis is used in trend
Figure 5.1
55
projections and one type of decomposition

model
56
28/01/16
Qualitative Models
Causal Models
n Qualitative models incorporate judgmental
n Causal models use variables or factors
or subjective factors
that might influence the quantity being

forecasted
n The objective is to build a model with
the best statistical relationship between
the variable being forecast and the
independent variables
n Regression analysis is the most
common technique used in causal
modeling
n Useful when subjective factors are
thought to be important or when accurate

quantitative data is difficult to obtain
n Common qualitative techniques are
n Delphi method
n Jury of executive opinion
n Sales force composite
n Consumer market surveys
57
Qualitative Models
58
Scatter Diagrams
Scatter diagrams are helpful when forecasting time-series
data because they depict the relationship between variables.
n Delphi Method an iterative group process where
Annual Sales
(possibly geographically dispersed) respondents

provide input to decision makers
n Jury of Executive Opinion collects opinions of a
small group of high-level managers, possibly
using statistical models for analysis
n Sales Force Composite individual salespersons
estimate the sales in their region and the data is
compiled at a district or national level
n Consumer Market Survey input is solicited from
customers or potential customers regarding their
purchasing plans
Radios
450
400
350
300
250
200
150
100
50
0
Televisions
Compact
0
Discs
10
12
Time (Years)
59
Scatter Diagrams
5 10
Scatter Diagrams
n Wacker Distributors wants to forecast sales for
three different products
1
2
3
4
5
6
7
8
9
10
Table 5.1
TELEVISION SETS
250
250
250
250
250
250
250
250
250
250
RADIOS
300
310
320
330
340
350
360
370
380
390
(a)
Annual Sales of Televisions
YEAR
COMPACT DISC PLAYERS

110
100
120
140
170
150
160
190
200
190
n Sales appear to be
330
250 l l l l l l l l l l
200
150
100
constant over time

Sales = 250
n A good estimate of
sales in year 11 is
250 televisions
50
|
0 1 2 3 4 5 6 7 8 9 10
Time (Years)
Figure 5.2
5 11
5 12
28/01/16
Scatter Diagrams
Annual Sales of Radios
(c)
420
increasing at a
constant rate of 10
radios per year
Sales = 290 + 10(Year)
n A reasonable
estimate of sales in
year 11 is 400
radios
400
380
360
340
320
300 l
280
|
0 1 2 3 4 5 6 7 8 9 10
n This trend line may

Annual Sales of CD Players
(b)
Scatter Diagrams
200
l
180
l
160
140
120
100
l
l
not be perfectly
accurate because
of variation from
year to year
increasing
n A forecast would
probably be a
larger figure each
year
l l
l
l
0 1 2 3 4 5 6 7 8 9 10
Time (Years)
Time (Years)
Figure 5.2
Figure 5.2
5 13
Measures of Forecast Accuracy
5 14

n Using a nave forecasting model
n We compare forecasted values with actual values
to see how well one model works or to compare

models
Forecast error = Actual value Forecast value
n One measure of accuracy is the mean absolute
deviation (MAD)
forecast error
MAD =
n
YEAR
ACTUAL
SALES OF CD
PLAYERS
FORECAST SALES
110
ABSOLUTE VALUE OF
ERRORS (DEVIATION),
(ACTUAL FORECAST)
100
110
|100 110| = 10
120
100
|120 100| = 20
140
120
|140 120| = 20
170
140
|170 140| = 30
150
170
|150 170| = 20
160
150
|160 150| = 10
190
160
|190 160| = 30
200
190
|200 190| = 10
10
190
200
11
190
|190 200| = 10
Sum of |errors| = 160

MAD = 160/9 = 17.8
5 15
YEAR
FORECAST SALES
ABSOLUTE VALUE OF
ERRORS (DEVIATION),
(ACTUAL FORECAST)
110
100
110
|100 110| = 10
120
100
|120 110| = 20
140
120
|140 120| = 20
170
140
|170 140| = 30
150
170
|150 170| = 20
160
150
|160 150| = 10
190
160
|190 160| = 30
200
190
|200 190| = 10
10
190
200
|190 200| = 10
11
190
accuracy
n The mean squared error
MSE =
Table 5.2
160
= 17 .8
9
error
MAPE =
actual
n
100%
n And bias is the average error and tells whether the
MAD = 160/9 = 17.8

(error )
n The mean absolute percent error
Sum of |errors| = 160

n
5 16
n There are other popular measures of forecast
ACTUAL
SALES OF CD
PLAYERS
forecast error
n Using a nave forecasting model
MAD =
Table 5.2
5 17
forecast tends to be too high or too low and by

how much. Thus, it can be negative or positive.
5 18
28/01/16
Hospital Days Forecast Error

Example

Year
Actual CD Sales
110
Forecast Sales
|Actual -Forecast|
100
110
10
120
100
20
140
120
20
170
140
30
150
170
20
160
150
10
190
160
30
200
190
10
10
190
200
10
11
190
Sum of |errors|
160
MAD
17.8
Forecast
250
Actual
243
FEB
320
315
MAR
275
286
APR
260
256
MAY
250
241
JUN
275
298
JUL
300
292
AUG
325
333
SEP
320
OCT
NOV
DEC
AVERAGE
|error|
error2
|error/actual|
49
0.03
25
0.02
11
121
0.04
16
0.02
81
0.04
23
529
0.08
64
0.03
64
0.02
326
36
0.02
350
378
28
784
0.07
365
382
17
289
380
396
16
MAD=
11.83
Actual
250
320
275
260
250
275
300
325
320
350
365
380
243
315
286
256
241
298
292
333
326
378
382
396
5 20
Time-Series Forecasting Models

n A time series is a sequence of evenly
spaced events (weekly, monthly, quarterly,

etc.)
n Time-series forecasts predict the future
based solely of the past values of the
variable
n Other variables, no matter how potentially
valuable, are ignored
0.04
256
MSE=
192.83
Forecast
5 19
Hospital Days Forecast Error

Example
JAN
Month
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
DEC
Ms. Smith forecasted

total hospital inpatient
days last year. Now
that the actual data are
known, she is
reevaluating her
forecasting model.
Compute the MAD,
MSE, and MAPE for her
forecast.
0.04
MAPE=
.0381*100 =
3.81
5 21
Decomposition of a Time-Series
Decomposition of a Time-Series
Demand for Product or Service
n A time series typically has four components
1.Trend (T) is the gradual upward or

downward movement of the data over time
2.Seasonality (S) is a pattern of demand
fluctuations above or below trend line that
repeats at regular intervals
3.Cycles (C) are patterns in annual data that
occur every several years
4.Random variations (R) are blips in the
data caused by chance and unusual
situations
Figure 5.3
5 22
5 23
Trend
Component
Seasonal Peaks
Actual
Demand
Line
Average Demand
over 4 Years
Year
1
Year
2
Time
Year
3
Year
4
5 24
28/01/16
Nave Forecast *
Nave Forecast *
n Nave forecast is the simplest technique. It
uses the actual demand for the past period as

the forecasted demand for the next period
n This makes the theory that the past will repeat.
n Also assumes that any time series
components are either reflected in the
previous periods demand or do not exist.
Nave forecast, Ft+1 = Yt
Actual Demand
35
40
35
55
40
65
55
60
65
60
5 26
Moving Averages
5 27
Moving Averages
n Moving averages can be used when demand is relatively steady
n Mathematically
over time
n The next forecast is the average of the most recent n data values
from the time series
n Conditions:
n There is a historical record of actual events.
n The forecast is based on a determined number of past
periods.
n The past periods have equal importance.
n The most recent period of data is added and the oldest is
dropped
nThis methods tends to smooth out short-term irregularities in
the data series
Moving average forecast =
Forecast
Period
Sum of demands in previous n periods

n
Ft +1 =
Yt + Yt 1 + ... + Yt n+1
n
where
Ft +1 = forecast for time period t + 1

Yt = actual value in time period t
n = number of periods to average
5 28
Wallace Garden Supply Example
5 29
n Wallace Garden Supply wants to
forecast demand for its Storage Shed

n They have collected data for the past
year
n They are using a three-month moving
average to forecast demand (n = 3)
MONTH
ACTUAL SHED SALES
January
10
THREE-MONTH MOVING AVERAGE
February
12
March
13
April
16
(10 + 12 + 13)/3 = 11.67
May
19
(12 + 13 + 16)/3 = 13.67
June
23
(13 + 16 + 19)/3 = 16.00
July
26
(16 + 19 + 23)/3 = 19.33
August
30
(19 + 23 + 26)/3 = 22.67
September
28
(23 + 26 + 30)/3 = 26.33
October
18
(26 + 30 + 28)/3 = 28.00
November
16
(30 + 28 + 18)/3 = 25.33
December
14
(28 + 18 + 16)/3 = 20.67
January
(18 + 16 + 14)/3 = 16.00
Table 5.3
5 30
5 31
28/01/16
Weighted Moving Averages
Weighted Moving Averages

n Both simple and weighted averages are
n Weighted moving averages use weights to put
more emphasis on recent periods

n Often used when a trend or other pattern is
emerging
Ft +1 =
effective in smoothing out fluctuations in

the demand pattern in order to provide
stable estimates
n Problems
( Weight in period i )( Actual value in period)

( Weights)
nIncreasing the size of n smoothes out
fluctuations better, but makes the method

less sensitive to real changes in the data
nMoving averages can not pick up trends
very well they will always stay within past
levels and not predict a change to a higher or
lower level
n Mathematically
Ft +1 =
w1Yt + w2Yt 1 + ... + w nYt n +1

w1 + w2 + ... + w n
where
wi = weight for the ith observation
5 32
n Wallace Garden Supply decides to try a
THREE-MONTH WEIGHTED
MOVING AVERAGE
MONTH
ACTUAL SHED SALES
January
10
February
12
March
13
April
16
[(3 X 13) + (2 X 12) + (10)]/6 = 12.17
May
19
[(3 X 16) + (2 X 13) + (12)]/6 = 14.33
June
23
[(3 X 19) + (2 X 16) + (13)]/6 = 17.00
July
26
[(3 X 23) + (2 X 19) + (16)]/6 = 20.50
August
30
[(3 X 26) + (2 X 23) + (19)]/6 = 23.83
September
28
[(3 X 30) + (2 X 26) + (23)]/6 = 27.50
October
18
[(3 X 28) + (2 X 30) + (26)]/6 = 28.33
3 x Sales last month + 2 x Sales two months ago + 1 X Sales three months ago
November
16
[(3 X 18) + (2 X 28) + (30)]/6 = 23.33
December
14
[(3 X 16) + (2 X 18) + (28)]/6 = 18.67
January
[(3 X 14) + (2 X 16) + (18)]/6 = 15.33
weighted moving average model to forecast

demand for its Storage Shed
n They decide on the following weighting
scheme
WEIGHTS APPLIED
PERIOD
3
2
1
Last month
Two months ago
Three months ago
5 33
Sum of the weights
Table 5.4
5 34
Exponential Smoothing
5 35
n Exponential smoothing is easy to use and
n Mathematically
requires little record keeping of data

n It is a type of moving average
n Conditions:
Ft +1 = Ft + (Yt Ft )
where
n There is historical data of actual sales.
Ft+1 = new forecast (for time period t + 1)

Ft = previous forecast (for time period t)
= smoothing constant (0 1)
Yt = previous periods actual demand
n There is a historical record of past forecasts.

n The smoothing constant is known.
New forecast =
Last periods forecast

+ (Last periods actual demand
Last periods forecast)
Where is a weight (or smoothing constant) with a
value between 0 and 1 inclusive
A larger gives more importance to recent data while
a smaller value gives more importance to past data
n The idea is simple the new estimate is the
old estimate plus some fraction of the error in

the last period
5 38
5 39
28/01/16
Selecting the Smoothing Constant
Exponential Smoothing Example

n In January, Februarys demand for a certain
n Selecting the appropriate value for
is
key to obtaining a good forecast
n The objective is always to generate an
accurate forecast
n The general approach is to develop trial
forecasts with different values of and
select the that results in the lowest MAD
car model was predicted to be 142
n Actual February demand was 153 autos

n Using a smoothing constant of
is the forecast for March?
= 0.20, what
New forecast (for March demand) = 142 + 0.2(153 142)

= 144.2 or 144 autos
n If actual demand in March was 136 autos, the
April forecast would be
New forecast (for April demand) = 144.2 + 0.2(136 144.2)

= 142.6 or 143 autos
5 40
Port of Baltimore Example

n Exponential smoothing forecast for two values of
QUARTER
ACTUAL
TONNAGE
UNLOADED
FORECAST
USING =0.10
5 41
Selecting the Best Value of

QUARTER
ACTUAL
TONNAGE
UNLOADED
180
175
168
175.5
FORECAST
USING =0.50
FORECAST
WITH = 0.10
180
175
175
168
175.5 = 175.00 + 0.10(180 175)
177.5
159
174.75
159
174.75 = 175.50 + 0.10(168 175.50)
172.75
175
173.18
175
173.18 = 174.75 + 0.10(159 174.75)
165.88
190
173.36 = 173.18 + 0.10(175 173.18)
170.44
190
173.36
205
175.02 = 173.36 + 0.10(190 173.36)
180.22
205
175.02
180
178.02 = 175.02 + 0.10(205 175.02)
192.61
180
178.02
182
178.22 = 178.02 + 0.10(180 178.02)
186.30
178.60 = 178.22 + 0.10(182 178.22)
184.15
182
Best choice
ABSOLUTE
DEVIATIONS
FOR = 0.10
5..
7.5..
15.75
1.82
16.64
29.98
1.98
178.22
3.78
Sum of absolute deviations
Table 5.5
5 42
Problem
MAD =
ABSOLUTE
DEVIATIONS
FOR = 0.50
175
5.
177.5
9.5..
172.75
13.75
165.88
9.12
170.44
19.56
180.22
24.78
192.61
12.61
186.30
4.3..
82.45
|deviations|
n
10.31
98.63
MAD =
12.33
5 43
PM Computer: Moving Average

Example
n PM Computer assembles customized personal
A firm uses simple exponential smoothing

with alpha = 0.1 to forecast demand. The
forecast for the week of January 1 was 500
units whereas actual demand turned out to
be 450 units.
computers from generic parts
n The owners purchase generic computer parts

in volume at a discount from a variety of
sources whenever they see a good deal.
What is the demand forecasted for the

week of January 8?
Table 5.6
FORECAST
WITH = 0.50
n It is important that they develop a good
forecast of demand for their computers so

they can purchase component parts
efficiently.
5 44
5 47
28/01/16
PM Computers: Moving Average

Solution
PM Computers: Data
Period
Month
Actual Demand
1
Jan
37
2
Feb
40
3
Mar
41
4
Apr
37
5
May
45
6
June
50
7
July
43
8
Aug
47
9
Sept
56
n
Compute a 2-month moving average
n
n
Abs. Dev
3 month WMA
Abs. Dev
Exp.Sm.
Compute a 3-month weighted average using weights of

4,2,1 for the past three months of data
Compute an exponential smoothing forecast using =
0.7, previous forecast of 40
Using MAD, what forecast is most accurate? 2009 Prentice-Hall, Inc.
MAD
37.00
3.00
39.10
1.90
3.14
40.43
3.43
38.57
6.43
38.03
6.97
9.00
42.14
7.86
42.91
7.09
47.50
4.50
46.71
3.71
47.87
4.87
46.50
0.50
45.29
1.71
44.46
2.54
45.00
11.00
46.29
9.71
46.24
9.76
38.50
2.50
40.50
3.50
40.14
39.00
6.00
41.00
51.50
51.57
53.07
5.29
5.43
4.95
Exponential smoothing resulted in the lowest MAD.

5 48
Trend Projection
5 49
Trend Projection
n Trend projections are used to forecast time-series data
n Trend projection fits a trend line to a
that exhibit a linear trend.

n A trend line is simply a linear regression equation
in which the independent variable (X) is the time
period
n The independent variable is the time period and the
dependent variable is the actual observed value in
the time series.
n Conditions:
n There is historical data of actual sales.
n There is an apparent linear trend in the figures
over time.
series of historical data points

n The line is projected into the future for
medium- to long-range forecasts
n Several trend equations can be
developed based on exponential or
quadratic models
n The simplest is a linear model developed
using regression analysis
5 53
Trend Projection
5 54
Trend Projection
Y = b + b1 X
n The mathematical form is:

0
Where:
Y = predicted value
b0 = intercept
b1 = slope of the line
X = time period (i.e., X = 1, 2, 3, , n)
b1 = xy nx y
------------x2 nx 2
Abs. Dev
37.00
Dist7
Value of Dependent Variable
2 month
MA
b0 = y b1 x
Where:
= summation sign for n data points
x = values of independent variables
y= values of dependent variables
n = number of data points or observations
x = average of the values of the xs
y = average of the values of the ys
*
Dist1
Dist5
Dist2
Time
5 55
Dist3
Dist6
Dist4
Figure 5.4
5 56
28/01/16
Midwestern Manufacturing
Company Example
Company Example
n Midwestern Manufacturing Company has
n The forecast equation is
experienced the following demand for its electrical

generators over the period of 2001 2007
YEAR
ELECTRICAL GENERATORS SOLD
2001
2002
2003
2004
2005
2006
2007
74
79
80
90
105
142
122
Y = 56.71+ 10.54 X
n To project demand for 2008, we use the coding
system to define X = 8
(sales in 2008) = 56.71 + 10.54(8)

= 141.03, or 141 generators
n Likewise for X = 9
(sales in 2009) = 56.71 + 10.54(9)

= 151.57, or 152 generators
Table 5.7
Determine the forecast for 2008 and 2009, and

plot a time series.
5 57
Company Example
Generator Demand
140
Trend Line
Y = 56.71+ 10.54 X
130
120
110
100
90
70
hrs/wk) over the past six

years has been:
150
80
Example 2
n The rated power capacity (in
160
l
l
Actual Demand Line
60
50
|
Figure 5.5
2001 2002 2003 2004 2005 2006 2007 2008 2009

Year
5 60
Alternative way to recode years

which simplifies math since x = 0
Renumbered
Yr., x
Capacity,
-2.5
115
6.25
-287.25
-1.5
120
2.25
-180
120
-0.5
118
0.25
-59
118
+0.5
124
0.25
+62
124
+1.5
123
2.25
+183.5
123
+2.5
130
6.25
+325
130
Year
Rated
Capacity
(hrs/wk)
115
Yr.
b1 = xy/x2= 5/17.5 = 2.57
730
X2
17.5
xy
45
b0 = y/n= 730/6 = 121.67
y = 121.67 + 2.57x
5 61
Seasonal Variations
5 64
Seasonal Variations
n Recurring variations over time may indicate the
n Eichler Supplies sells telephone
need for seasonal adjustments in the trend line
answering machines
n A seasonal index indicates how a particular
n Data has been collected for the past two
season compares with an average season

n When no trend is present, the seasonal index
can be found by dividing the average value for
a particular season by the average of all the
data
n Conditions:
years sales of one particular model
n They want to create a forecast that
includes seasonality
n There is historical record of actual events.

n The past record of actual events is at least 2 years.
n There is an apparent trend in the figures over time.
n There is an apparent trend in the figures across
seasons.
5 65
5 66
28/01/16
Seasonal Variations
SALES DEMAND
MONTH
YEAR 1
YEAR 2
January
80
100
February
85
75
March
80
90
110
90
May
115
131
June
120
110
July
100
110
August
110
90
September
85
95
October
75
85
November
85
75
December
80
80
April
AVERAGE TWO- YEAR

DEMAND
90
80
85
100
123
115
105
100
90
80
80
80
Seasonal Variations
MONTHLY
DEMAND
AVERAGE
SEASONAL INDEX
94
0.957
94
0.851
94
0.904
94
1.064
94
1.309
94
1.223
94
1.117
94
1.064
94
0.957
94
0.851
94
0.851
94
0.851
n Suppose we expected the third years annual demand for
answering machines to be 1,200 units, which is 100 per

month. We would not forecast each month to have a demand
of 100, but we would ad- just these based on the seasonal
indices as follows:
Total average demand = 1,128
Average monthly demand =
1,128
= 94
12 months
Average two-year demand

Seasonal index =
Average monthly demand
Table 5.8
Jan.
1,200
0.957 = 96
12
July
1,200
1.117 = 112
12
Feb.
1,200
0.851 = 85
12
Aug.
1,200
1.064 = 106
12
Mar.
1,200
0.904 = 90
12
Sept.
1,200
0.957 = 96
12
Apr.
1,200
1.064 = 106
12
Oct.
1,200
0.851 = 85
12
May
1,200
1.309 = 131
12
Nov.
1,200
0.851 = 85
12
June
1,200
1.223 = 122
12
Dec.
1,200
0.851 = 85
12
5 67
174
5 68
CHAPTER 5 FORECASTING
TABLE 5.11
Centered Moving
Averages and
Seasonal Ratios for
Turner Industries
YEAR
QUARTER
SALES ($1,000,000s)
108
125
150
CMA
132.000
SEASONAL RATIO
1.136
141
134.125
1.051
116
136.375
0.851
134
138.875
159
141.125
0.965
1.127
Steps Used to Compute Seasonal

Indices based on CMAs
Seasonal Variations with Trends
n When both trend and seasonal components are present in a time
series, a change from one month to the next could be due to a

trend, to a seasonal variation, or simply to random fluctuations.
n To help with this problem, the seasonal indices should be
computed using centered moving average approach whenever
trend is present.
n Using this approach prevents a variation due to trend from being
incorrectly interpreted as a variation due to the season.
n Conditions:
n There is historical record of actual events.
152
143.000
1.063
123
145.125
0.848
142
147.875
0.960
168
165
ter 1 for
for year 2.each
The average will
be
1. Compute a CMA
observation
(where
we take quarters 2, 3, and 4 of year 1, plus one-half of quarter 1 for year 1 and one-half of quar-
0.511082 + 125 + 150 + 141 + 0.511162

possible)
= 132.00
CMA 1quarter 3 of year 12 =
4
We compare the actual sales in this quarter to the CMA and we have the following seasonal
2. Compute seasonal
ratio = Observation/CMA for
ratio:
Sales in quarter 3
150
that observation.
=
= 1.136
Seasonal ratio =
CMA
132.00
Thus, sales
in quarter 3 of to
year 1 are
about 13.6%
higher than an average
quarter at this time.
3. Average seasonal
ratios
get
seasonal
indices.
All of the CMAs and the seasonal ratios are shown in Table 5.11.
Since
there
are
two
seasonal
ratios
for
each
quarter,
we
average
these to get the seasonal
(This eliminatesindex.as
much randomness as
Thus,
possible.)
Index for quarter 1 = I = 10.851 + 0.8482>2 = 0.85
4. If seasonal indices do Index
not
add
the
number
of
= 11.136
+ 1.1272>2
= 1.13
for quarter
3 = I to
seasons, multiply
each index by (Number of
The sum of these indices should be the number of seasons (4) since an average season should
have an index of 1. In this example, the sum is 4. If the sum were not 4, an adjustment would be
seasons)/(Summade.
ofWethe
indices).
would multiply
each index by 4 and divide this by the sum of the indices.
1
2
3
4
n The past records of actual events is at least 2 years.

n There is an apparent trend in the figures over time.
n There is an apparent trend in the figures across seasons.
Steps Used to Compute Seasonal Indices Based on CMAs
n There is a need to deseasonalize the projection for accuracy.
1.
2.
3.
4.
5 69
Compute a CMA for each observation (where possible).

Compute seasonal ratio = Observation/CMA for that observation.
Average seasonal ratios to get seasonal indices.
If seasonal indices do not add to the number of seasons, multiply each index by (Number
of seasons)/(Sum of the indices).
5 70
Figure 5.5 provides a scatterplot of the Turner Industries data and the CMAs. Notice that
the plot of the CMAs is much smoother than the original data. A definite trend is apparent in the
data.
Turner Industries Example
Consider Q3. The actual sales in that quarter

were 150. To determine the magnitude of the
seasonal variation, we should compare this
with an average quarter centered at that time
period. Thus, we should have a total of four
quarters (1 year of data) with an equal number
of quarters before and after quarter 3 so the
trend is averaged out. Thus, we need 1.5
quarters before quarter 3 and 1.5 quarters
after it. To obtain the CMA, we take quarters 2,
3, and 4 of year 1, plus one-half of quarter 1
for year 1 and one-half of quarter 1 for year 2.
Seasonal Ratio =
observation/CMA for
that observation
5 71
Average seasonal ratios to

get seasonal
indices.
2009 Prentice-Hall, Inc. 5 72
10
28/01/16
Scatter Plot of Turner

Industries Sales and CMA
Seasonal Ratio =
observation/CMA for
that observation
Average seasonal ratios to
get seasonal
indices.
2009 Prentice-Hall, Inc. 5 73
5 74
Steps to Develop a Forecast Using

the Decomposition Method
The Decomposition Method with

Trend and Seasonal Components
n Decomposition is the process of isolating linear
1. Compute seasonal indices using CMAs.
trend and seasonal factors to develop more

accurate forecasts
n The first step is to compute seasonal indices for
each season, then the data are deseasonalized
by dividing each number by its seasonal index
n A trend line is then found using the
deseasonalized data.
2. Deseasonalize the data by dividing each number
by its seasonal index.
3. Find the equation of a trend line using the
deseasonalized data.
4. Forecast the future periods using the trend line.

5. Multiply the trend line forecast by the
appropriate seasonal index
174
5 75
5 76
CHAPTER 5 FORECASTING
TABLE 5.11
Centered Moving
Averages and
Seasonal Ratios for
Turner Industries
YEAR
QUARTER
SALES ($1,000,000s)
108
125
CMA
SEASONAL RATIO
Deseasonalized Data for

Turner Industries
3
150
132.000
1.136
141
134.125
1.051
116
136.375
0.851
134
138.875
0.965
159
141.125
1.127
152
143.000
1.063
123
145.125
0.848
142
147.875
0.960
168
165
Finding the Trend Line of

Deseasonalized Data
b1 = 2.34, b0 = 124.78
Y = 124.78 + 2.34X where X = time
we take quarters 2, 3, and 4 of year 1, plus one-half of quarter 1 for year 1 and one-half of quarter 1 for year 2. The average will be
CMA 1quarter 3 of year 12 =
0.511082 + 125 + 150 + 141 + 0.511162

4
This equation is used to develop the forecast based on

trend, and the result is multiplied by the appropriate seasonal
index to make a seasonal adjustment.
The forecast for the first quarter of year 4 (time period = 13
and seasonal index I1 = 0.85)
= 132.00
We compare the actual sales in this quarter to the CMA and we have the following seasonal
ratio:
Seasonal ratio =
Sales in quarter 3
150
=
= 1.136
CMA
132.00
Thus, sales in quarter 3 of year 1 are about 13.6% higher than an average quarter at this time.
All of the CMAs and the seasonal ratios are shown in Table 5.11.
Since there are two seasonal ratios for each quarter, we average these to get the seasonal
index. Thus,
Y = 124.78 + 2.34X = 124.78 + 2.34(13)

= 155.2 (forecast before adjustment for seasonality)
Index for quarter 1 = I1 = 10.851 + 0.8482>2 = 0.85

Multiply this by the seasonal index for quarter 1:
The sum of these indices should be the number of seasons (4) since an average season should
have an index of 1. In this example, the sum is 4. If the sum were not 4, an adjustment would be
made. We would multiply each index by 4 and divide this by the sum of the indices.
Y x I1 = 155.2 x 0.85 = 131.92
Find the forecast for quarters 2, 3 and 4 of the next year.
Steps Used to Compute Seasonal Indices Based on CMAs

1.
2.
3.
4.

Compute a CMA for each observation (where possible).
Compute seasonal ratio = Observation/CMA for that observation.
Average seasonal ratios to get seasonal indices.
If seasonal indices do not add to the number of seasons, multiply each index by (Number
of seasons)/(Sum of the indices).
5 77
5 78
Figure 5.5 provides a scatterplot of the Turner Industries data and the CMAs. Notice that
the plot of the CMAs is much smoother than the original data. A definite trend is apparent in the
data.
11
28/01/16
Example 2
n Demand for Carriers most
popular window air

conditioner units is quite
seasonal. For each quarter
of the past three years,
demand has been as
follows:
n Find the seasonal indices
for aircon demand.
n Use the seasoned indices
developed and
deseasonalized the 12
quarters of demand
Example 3
Period
Year
Quarter
Demand
1986
126
200
243
167
132
211
243
167
131
10
208
11
251
12
171
1987
1988
Month
Sales
Last Year
n Del Monte Catsup wants to
improve the accuracy of the

forecast by removing the
effect of seasonal trend
based on month in the trend
line formula. There is a linear
trend of the monthly sales
over time with variation due
to the seasons based on
month over the last 2 years.
What is the forecast of the
monthly sales next year?
5 79
This Year
Jan
Feb
11
15
Mar
15
17
Apr
16
16
May
15
16
Jun
25
23
Jul
21
20
Aug
24
25
Sep
24
28
Oct
19
21
Nov
20
Dec
25
28
29
5 80
MS Application-Forecasting:
Taco Bell
Example 3 [Steps]
n Determine the seasonal index.
n Determine the deseasonalized sales.
n Determine the trend line formula.
n Determine the adjusted monthly
sales forecast for next year.
5 81
5 93
Question of the Day
Using The Computer to Forecast

n Spreadsheets can be used by small and
medium-sized forecasting problems
n More advanced programs (SAS, SPSS,
Minitab) handle time-series and causal

models
n May automatically select best model
parameters
n Dedicated forecasting packages may be
fully automatic
n May be integrated with inventory planning
and control
5 94
5 95
12

Notes On Forecasting

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Notes On Forecasting

Încărcat de

Drepturi de autor:

Formate disponibile

28/01/16

n Managers are always trying to reduce

uncertainty and make better estimates of what

2008 Prentice-Hall, Inc.

2009 Prentice-Hall, Inc.

n Eight steps to forecasting :

n These steps are a systematic way of initiating,

1. Determine the use of the forecastwhat

2009 Prentice-Hall, Inc.

designing, and implementing a forecasting

n Time-series models attempt to predict the

future based on the past

n Common time-series models are

2009 Prentice-Hall, Inc.

projections and one type of decomposition

n Qualitative models incorporate judgmental

n Causal models use variables or factors

that might influence the quantity being

n Useful when subjective factors are

thought to be important or when accurate

2009 Prentice-Hall, Inc.

n Delphi Method an iterative group process where

2009 Prentice-Hall, Inc.

(possibly geographically dispersed) respondents

2009 Prentice-Hall, Inc.

n Wacker Distributors wants to forecast sales for

three different products

COMPACT DISC PLAYERS

constant over time

2009 Prentice-Hall, Inc.

Annual Sales of Radios

n This trend line may

Measures of Forecast Accuracy

2009 Prentice-Hall, Inc.

Measures of Forecast Accuracy

n We compare forecasted values with actual values

to see how well one model works or to compare

n One measure of accuracy is the mean absolute

Sum of |errors| = 160

2009 Prentice-Hall, Inc.

Measures of Forecast Accuracy

n The mean squared error

n And bias is the average error and tells whether the

MAD = 160/9 = 17.8

n The mean absolute percent error

Sum of |errors| = 160

n There are other popular measures of forecast

2009 Prentice-Hall, Inc.

Measures of Forecast Accuracy

n Using a nave forecasting model

forecast tends to be too high or too low and by

2009 Prentice-Hall, Inc.

Hospital Days Forecast Error

Measures of Forecast Accuracy

2009 Prentice-Hall, Inc.

2009 Prentice-Hall, Inc.

Time-Series Forecasting Models

spaced events (weekly, monthly, quarterly,

Hospital Days Forecast Error

Ms. Smith forecasted

2009 Prentice-Hall, Inc.

n A time series typically has four components

1.Trend (T) is the gradual upward or