Chapter 15

Forecasting
Chapter 15
Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-1

Chapter Topics
■ Forecasting
■ Forecasting Components
■ Time Series Methods
■ Forecast Accuracy

Forecasting
A forecast is a prediction of what will occur in the future
Examples
• Meteorologists forecast the weather,
• sportscasters predict the winners of football games, and
• managers of business firms attempt to predict how much of their
product will be demanded in the future.
 Good forecasts can lead to
 Reduced inventory costs.

 Lower overall personnel costs.
 Increased customer satisfaction.

Forecasting Components
■ A variety of forecasting methods are available for use depending on the

time frame of the forecast and the existence of patterns.
■ Time Frames:
 Short-range (one to two months) such as daily demand or resource
requirements.
 Medium-range (two months to one or two years) additional
resources for the upcoming year.
 Long-range (more than one or two years) new products for
changing markets, build new facilities.
■ Patterns:
 Trend
 Random variations
 Cycles
 Seasonal pattern
Patterns (1 of 2)
 Trend - A long-term movement of the item being forecast.
For example, the demand for personal computers has shown an upward
trend during the past decade,
 Random variations - movements that are not predictable and follow no
pattern.
 Cycle - A movement, up or down, that repeats itself over a lengthy
time span.
For example, The demand for winter sports equipment increases every 4
years, before and after the Winter Olympics
 Seasonal pattern - Oscillating movement in demand that occurs
periodically in the short run.

Patterns (2 of 2)
 For example, retail sales in general increase during the Christmas season.
However, a seasonal pattern can occur on a daily or weekly basis. some
restaurants are busier at lunch than at dinner, and shopping mall stores and
theaters tend to have higher demand on weekends.
Figure 15.1 (a) Trend; (b) Cycle; (c) Seasonal; (d) Trend with Season
Forecasting Methods
1. Times Series - Statistical techniques that use historical

data to predict future behavior.
2. Regression Methods - Regression (or causal ) methods
that attempt to develop a mathematical relationship
between the item being forecast and factors that cause it
to behave the way it does.
3. Qualitative Methods - Methods using judgment, expertise
and opinion to make forecasts.

Time Series Methods
Overview
 Statistical techniques that make use of historical data collected

over a long period of time.
 Methods assume that what has occurred in the past will

continue to occur in the future.
 Forecasts based on only one factor - time.

Time Series Methods
Moving Average (1 of 6)
 Moving average uses values from the recent past to develop
forecasts.
 This dampens random increases and decreases.
 Useful for forecasting relatively stable items that do not display
any trend or seasonal pattern.
 Moving averages are computed for specific periods, such as 3
months or 5 months, depending on how much the forecaster desires
to smooth the data.
 Formula for moving average (MA):
n
 Di
MAn  i1n
where:
n  number of periods in the moving average
D  data in period i
i
Time Series Methods
Example: Instant Paper Clip Supply Company wants to forecast
orders for the month of November. Develop three-month and
five-month moving averages using the data.
Table 15.1 Orders for 10-month period

Time Series Methods
Example: Instant Paper Clip Supply Company wants to forecast
orders for the month of November.
 Three-month moving average:
3
 Di
MA  i1  90 110 130 110 orders
3 3 3
 Five-month moving average:

5
 Di
MA  i1  90 110 130  75  50  91 orders
5 5 5

Time Series Methods
Table 15.2 Three- and 5-month moving averages

Time Series Methods
Figure 15.2 Three- and 5-month moving averages

Time Series Methods
 Both moving average forecasts in Table 15.2 tend to smooth out the
variability occurring in the actual data.
 Notice that the 5-month moving average in Figure 15.2 smoothes
out fluctuations to a greater extent than the 3-month moving
average.
 The appropriate number of periods to use often requires trial-and-
error experimentation.
 The major disadvantage of moving average does not react well to
changes (trends, seasonal effects, etc.)
 but is easy to use, quick and inexpensive.
 Good for short-term forecasting.

Time Series Methods
Weighted Moving Average
 In a weighted moving average, weights are assigned to the most
recent data.
n
WMA   W D
n i1 i i
where W  the weight for period i, between 0% and 100%

i
Wi  1.00
Example: Paper clip company weights 50% for October, 33%

for September, 17% for August:
3
WMA   W D  (.50)(90)  (.33)(110)  (.17)(130)  103.4 orders
i i
3 i1
 Determining precise weights and the number of periods requires
trial-and-error experimentation.
Time Series Methods
Exponential Smoothing (1 of 11)
 Exponential smoothing weights recent past data more strongly

than more distant data.
 Two forms: simple exponential smoothing and adjusted
exponential smoothing.
 Simple exponential smoothing:
Ft + 1 = Dt + (1 - )Ft
where:
Ft + 1 = the forecast for the next period
Dt = actual demand in the present period
Ft = the previously determined forecast for the present period
 = a weighting factor (smoothing constant).

Time Series Methods
 The smoothing constant, is between zero and one. It
reflects the weight given to the most recent demand data.
For example, if Ft + 1 = 0.2 Dt + 0.8 Ft

which means that our forecast for the next period is based
on 20% of recent demand and 80% of forecasted (past)
demand
 If we go to one extreme and let  = 0.0, then

Ft + 1= Ft
 In the other hand, if =1 then
Ft + 1= Dt
Time Series Methods
 The most commonly used values of  are between 0.10 and

0.50.
 Determination of  is usually judgmental and subjective

and often based on trial-and -error experimentation.

Time Series Methods
Example: PM Computer Services (see Table 15.4).

 Exponential smoothing forecasts using smoothing constant of .30.
 Forecast for period 2 (February):

F2 =  D1 + (1- )F1 = (.30)(.37) + (1-.30)(.37) = 37 units
 Forecast for period 3 (March):

F3 =  D2 + (1- )F2 = (.30)(.40) + (1-.30)(37) = 37.9 units

Time Series Methods
Table 15.4 Exponential smoothing forecasts,  = .30 and  = .50

Time Series Methods
 The forecast that uses the higher smoothing constant (.50) reacts
more strongly to changes in demand than does the forecast with the
lower constant (.30).
 Both forecasts lag behind actual demand.
 Both forecasts tend to be consistently lower than actual demand.
 In general, when demand is relatively stable, without any trend,
using a small value for  is more appropriate to simply smooth out
the forecast.
 Alternatively, when actual demand displays an increasing (or
decreasing) trend, as is the case in Figure 15.3, a larger value of  is
generally better.

Time Series Methods
Figure 15.3 Exponential smoothing forecasts

Time Series Methods
■ Adjusted exponential smoothing: exponential smoothing with a

trend adjustment factor added.
Formula AFt + 1 = Ft + 1 + Tt+1
where:
T = an exponentially smoothed trend factor
Tt + 1 = (Ft + 1 - Ft) + (1 - )Tt
Tt = the last period trend factor
 = smoothing constant for trend ( a value between zero and one).
■ Reflects the weight given to the most recent trend data.
■ Determined subjectively.
■ A high  reflects trend changes more than a low  .
■ It is not uncommon for  to equal  in this method.
Time Series Methods
Example: PM Computer Services exponentially smoothed

forecasts with  = .50 and  = .30 (see Table 15.5 next slide).
The adjusted forecast for February AF2 , is the same as the
exponentially smoothed forecast because the trend computing
factor will be zero (i.e F1 and F2 are the same and T2 =0)
Adjusted forecast for period 3:
T3 = (F3 - F2) + (1 - )T2
= (.30)(38.5 - 37.0) + (.70)(0) = 0.45
AF3 = F3 + T3 = 38.5 + 0.45 = 38.95

Time Series Methods
Table 15.5 Adjusted exponentially smoothed forecast values

Time Series Methods
■ The adjusted forecast is consistently higher than the simple exponentially smoothed
forecast.
■ It is more reflective of the generally increasing trend of the data.
Figure 15.4 Adjusted exponentially smoothed forecast

Time Series Methods
Linear Trend Line (1 of 5)
■ When demand displays an obvious trend over time, a least squares
regression line , or linear trend line, can be used to forecast.
■ Formula:
y  a  bx b   xy  nxy
 x2  nx
where: a  y bx
a  intercept (at period 0)
b  slope of the line where:
x  the time period n  number of periods
y  forecast for demand x  nx
for period x
y  ny

Time Series Methods
Example: PM Computer Services (see Table 15.6)
x  78  6.5 y  557  46.42

12 12
b  xy  nxy  3,867  (12)(6.5)(46.42) 1.72

 x 2  nx2 650 12  6.52
a  y  bx  46.42  (1.72)(6.5)  35.2

y  35.2 1.72x linear trend line
for period 13, x  13, y  35.2 1.72(13)  57.56

Time Series Methods
Table 15.6 Least squares calculations

Time Series Methods
■ A trend line does not adjust to a change in the trend as does the
exponential smoothing method, it is assumed that all future
forecasts will follow a straight line
■ This limits its use to shorter time frames in which the trend will
not change.

Time Series Methods
Linear Trend (5 of 5)
Figure 15.5 Linear trend line

Time Series Methods
Seasonal Adjustments (1 of 4)
■ A seasonal pattern is a repetitive up-and-down movement in

demand.
■ Seasonal patterns can occur on a quarterly, monthly, weekly, or daily
basis.
■ A seasonally adjusted forecast can be developed by multiplying
the normal forecast by a seasonal factor.
■ A seasonal factor can be determined by dividing the actual
demand for each seasonal period by total annual demand:
Si =Di/D

Time Series Methods
■ Seasonal factors lie between zero and one and represent the portion
of total annual demand assigned to each season.
■ Seasonal factors are multiplied by annual demand to provide

adjusted forecasts for each period.

Time Series Methods
Example: Wishbone Farms
Table 15.7 Demand for turkeys at Wishbone Farms

S1 = D1/ D = 42.0/148.7 = 0.28
S2 = D2/ D = 29.5/148.7 = 0.20
S3 = D3/ D = 21.9/148.7 = 0.15
S4 = D4/ D = 55.3/148.7 = 0.37
Time Series Methods
 Multiply forecasted demand for an entire year by seasonal factors to

determine the quarterly demand.
 Forecast for entire year (trend line for data in Table 15.7):
y = 40.97 + 4.30x = 40.97 + 4.30(4) = 58.17
 Seasonally adjusted forecasts:
SF1 = (S1)(F5) = (.28)(58.17) = 16.28
SF2 = (S2)(F5) = (.20)(58.17) = 11.63
SF3 = (S3)(F5) = (.15)(58.17) = 8.73
SF4 = (S4)(F5) = (.37)(58.17) = 21.53

Forecast Accuracy
Overview
 Forecasts will always deviate from actual values.
 Difference between forecasts and actual values are referred to as
forecast error.
 We would like forecast error to be as small as possible.
 If forecast error is large, either the technique being used is the
wrong one, or the parameters need adjusting.
 Measures of forecast errors:
 Mean Absolute deviation (MAD)
 Mean absolute percentage deviation (MAPD)
 Cumulative error (E bar)
 Average error, or bias (E)

Forecast Accuracy
Mean Absolute Deviation (1 of 6)
 MAD is the average absolute difference between the forecast
and actual demand.
 The most popular and simplest-to-use measures of forecast error.
 Formula:
 Dt  Ft
MAD  n
where:
t  the period number
D  demand in period t
t
F  the forecast for period t
t
n  the total number of periods

Forecast Accuracy
Example: PM Computer Services (see Table 15.8).
Compare accuracies of different forecasts using MAD:
 Dt  Ft 53.41
MAD  n   4.85
11

Forecast Accuracy
Table 15.8 Computational values for MAD and error

Forecast Accuracy
 The lower the value of MAD relative to the magnitude of the

data, the more accurate the forecast.
 Can be used to compare the accuracy of different forecasting
techniques working on the same set of demand data (PM Computer
Services):
Exponential smoothing ( = .50): MAD = 4.04
Adjusted exponential smoothing ( = .50,  = .30): MAD = 3.81
Linear trend line: MAD = 2.29
 The linear trend line has the lowest MAD; increasing  from .30 to
.50 improved the smoothed forecast.
Forecast Accuracy
 A variation on MAD is the mean absolute percent deviation

(MAPD).
 Measures the absolute error as a percentage of demand rather
than per period.
 Eliminates the problem of interpreting the measure of accuracy
relative to the magnitude of the demand and forecast values.
 Formula:
 Dt  Ft 53.41
MAPD   .103 or 10.3%
 Dt 520

Forecast Accuracy
MAPD for other three forecasts:

Exponential smoothing ( = .50): MAPD = 8.5%
Adjusted exponential smoothing ( = .50,  = .30):
MAPD = 8.1%
Linear trend: MAPD = 4.9%

Forecast Accuracy
Cumulative Error (1 of 2)
 Cumulative error is the sum of the forecast errors (E =et).
 A relatively large positive value indicates the forecast is biased
low, a large negative value indicates the forecast is biased high.
 If the preponderance of errors are positive, the forecast is
consistently low; and vice versa.
 The cumulative error for a trend line is always almost zero, and
is therefore not a good measure for this method.
 The cumulative error for PM Computer Services can be read
directly from Table 15.8.
 E =  et = 49.31, indicating the forecasts are frequently below
actual demand.

Forecast Accuracy
Cumulative Error (2 of 2)
 Cumulative error for other forecasts:
Exponential smoothing ( = .50): E = 33.21
Adjusted exponential smoothing ( = .50,  =.30):
E = 21.14
 Average error (bias) is the per-period average of cumulative error.
 Average error for the exponential smoothing forecast:

E n 
et 49.31
 4.48
11
 A large positive value of average error indicates a forecast is biased
low; a large negative error indicates it is biased high.

Forecast Accuracy
Example Forecasts by Different Measures
Table 15.9 Comparison of forecasts for PM Computer Services

Results consistent for all forecasts:
 Larger value of alpha is preferable.
 Adjusted forecast is more accurate than exponential smoothing.
 Linear trend is more accurate than all the others.

Chapter 15

Uploaded by

Copyright:

Available Formats

You might also like

Chapter 15

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chapter 15

Uploaded by

Copyright:

Available Formats

Forecasting

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-1

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-2

A forecast is a prediction of what will occur in the future

 Good forecasts can lead to

 Reduced inventory costs.

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-3

■ A variety of forecasting methods are available for use depending on the

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-5

1. Times Series - Statistical techniques that use historical

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-7

 Statistical techniques that make use of historical data collected

 Methods assume that what has occurred in the past will

 Forecasts based on only one factor - time.

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-8

Table 15.1 Orders for 10-month period

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-10

 Five-month moving average:

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-11

Table 15.2 Three- and 5-month moving averages

Figure 15.2 Three- and 5-month moving averages

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-14

where W  the weight for period i, between 0% and 100%

Example: Paper clip company weights 50% for October, 33%

 Exponential smoothing weights recent past data more strongly

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-16

For example, if Ft + 1 = 0.2 Dt + 0.8 Ft

 If we go to one extreme and let  = 0.0, then

 The most commonly used values of  are between 0.10 and

 Determination of  is usually judgmental and subjective

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-18

Example: PM Computer Services (see Table 15.4).

 Forecast for period 2 (February):

 Forecast for period 3 (March):

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-19

Table 15.4 Exponential smoothing forecasts,  = .30 and  = .50

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-21

Figure 15.3 Exponential smoothing forecasts

■ Adjusted exponential smoothing: exponential smoothing with a

Example: PM Computer Services exponentially smoothed

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-24

Table 15.5 Adjusted exponentially smoothed forecast values

Figure 15.4 Adjusted exponentially smoothed forecast

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-27

Example: PM Computer Services (see Table 15.6)

x  78  6.5 y  557  46.42

b  xy  nxy  3,867  (12)(6.5)(46.42) 1.72

a  y  bx  46.42  (1.72)(6.5)  35.2

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-28

Table 15.6 Least squares calculations

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-30

Figure 15.5 Linear trend line

■ A seasonal pattern is a repetitive up-and-down movement in

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-32

■ Seasonal factors are multiplied by annual demand to provide

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-33

Table 15.7 Demand for turkeys at Wishbone Farms

 Multiply forecasted demand for an entire year by seasonal factors to

Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall 15-35