Professional Documents
Culture Documents
Forecasting
McGraw-Hill/Irwin
Copyright © 2012 by The McGraw-Hill Companies, Inc. All rights reserved.
Chapter 3: Learning Objectives
You should be able to:
1. List the elements of a good forecast
2. Outline the steps in the forecasting process
3. Describe at least three qualitative forecasting techniques and the
advantages and disadvantages of each
4. Compare and contrast qualitative and quantitative approaches to
forecasting
5. Describe averaging techniques, trend and seasonal techniques,
and regression analysis, and solve typical problems
6. Explain three measures of forecast accuracy
7. Compare two ways of evaluating and controlling forecasts
8. Assess the major factors and trade-offs to consider when
choosing a forecasting technique
#profevad2015
2
Forecast
Forecast – a statement about the future value of a
variable of interest
We make forecasts about such things as weather,
demand, and resource availability
Forecasts are an important element in making informed
decisions
#profevad2015
3
Forecasts affect decisions and activities throughout an organization
#profevad2015
5
Features Common to All Forecasts
1. Techniques assume some underlying causal system that
existed in the past will persist into the future
2. Forecasts are not perfect
3. Forecasts for groups of items are more accurate than
those for individual items
4. Forecast accuracy decreases as the forecasting horizon
increases
#profevad2015
6
Elements of a Good Forecast
The forecast
should be timely
should be accurate
should be reliable
should be expressed in meaningful units
should be in writing
technique should be simple to understand and use
should be cost effective
#profevad2015
7
Steps in the Forecasting Process
1. Determine the purpose of the forecast
2. Establish a time horizon
3. Obtain, clean, and analyze appropriate data
4. Select a forecasting technique
5. Make the forecast
6. Monitor the forecast
#profevad2015
8
Forecasting Approaches
Qualitative Forecasting
Qualitative techniques permit the inclusion of soft information such as:
Human factors
Personal opinions
Hunches
#profevad2015 9
Judgmental Forecasts
Forecasts that use subjective inputs such as opinions
from consumer surveys, sales staff, managers,
executives, and experts
Executive opinions
Salesforce opinions
Consumer surveys
Delphi method
#profevad2015 10
Time-Series Forecasts
Forecasts that project patterns identified in recent
time-series observations
Time-series - a time-ordered sequence of observations
taken at regular time intervals
Assume that future values of the time-series can be
estimated from past values of the time-series
#profevad2015 3-11
11
Time-Series Behaviors
Trend
Seasonality
Cycles
Irregular variations
Random variation
#profevad2015 3-12
12
Trends and Seasonality
Trend
A long-term upward or downward movement in data
Population shifts
Changing income
Seasonality
Short-term, fairly regular variations related to the calendar or time
of day
Restaurants, service call centers, and theaters all experience
seasonal demand
#profevad2015 13
Cycles and Variations
Cycle
Wavelike variations lasting more than one year
These are often related to a variety of economic, political, or even
agricultural conditions
Random Variation
Residual variation that remains after all other behaviors have been
accounted for
Irregular variation
Due to unusual circumstances that do not reflect typical behavior
Labor strike
Weather event
#profevad2015 14
Time-Series Behaviors
#profevad2015
15
Time-Series Forecasting - Naïve Forecast
Naïve Forecast
Uses a single previous value of a time series as the basis
for a forecast
The forecast for a time period is equal to the previous
#profevad2015
16
Naïve Forecasts
Forecast for any period = previous period’s actual
value
Ft = At-1
F: forecast A: Actual t: time period
#profevad2015
Naïve Forecast Example
#profevad2015
Naïve Forecasts
Simple to use
Virtually no cost
Quick and easy to prepare
Data analysis is nonexistent
Easily understandable
Cannot provide high accuracy
Can be a standard for accuracy
#profevad2015
Uses for Naïve Forecasts
#profevad2015
Time-Series Forecasting - Averaging
#profevad2015
21
Moving Average
Technique that averages a number of the most recent
actual values in generating a forecast
n
A t i
Ft MA n i 1
n
where
Ft Forecast for time period t
MA n n period moving average
At 1 Actual value in period t 1
n Number of periods in the moving average
#profevad2015
22
Moving Average
As new data become available, the forecast is
updated by adding the newest value and dropping the
oldest and then re-computing the average
The number of data points included in the average
determines the model’s sensitivity
Fewer data points used-- more responsive
More data points used-- less responsive
#profevad2015
23
Moving Average Example
t A F = MA3 A-F
20
1 -
2 25 -
3 15 -
4 30 20 10
5 27 23.3333 3.66667
6 24
#profevad2015
24
Simple Moving Average
Figure 3-4 Revised Forecast Forecast
Actual (MA3) (MA5)
47
45
43
41
39
37
35
1 2 3 4 5 6 7 8 9 10 11 12
Questions:
• Why is MA3 longer than MA5?
• Which curve fluctuate the most?
• Which curve is the smoothest?
#profevad2015
Simple Moving Average
Responsiveness vs. Stability
• Smaller m, responsiveness , stability
• Larger m, responsiveness , stability
• Must maintain stability when fluctuations are high.
Figure 3-4 Revised Forecast Forecast
Actual (MA3) (MA5)
47
45
43
41
39
37
35
1 2 3 4 5 6 7 8 9 10 11 12
#profevad2015
Weighted Moving Average
The most recent values in a time series are given more
weight in computing a forecast
The choice of weights, w, is somewhat arbitrary and
involves some trial and error
Ft wt ( At ) wt 1 ( At 1 ) ... wt n ( At n )
where
wt weight for period t , wt 1 weight for period t 1, etc.
At the actual value for period t , At 1 the actual value for period t 1, etc.
#profevad2015
27
Weighted Moving Average Example
Week Sales (actual) Sales (forecast) Error
t A F = MA3 A-F
1 20 -
2 25 -
3 15 -
4 30 19 11
5 27 24.5 2.5
6 25.5
#profevad2015
Exponential Smoothing
A weighted averaging method that is based on the
previous forecast plus a percentage of the forecast
Ft Ft 1 ( At 1 Ft 1 )
error
where
Ft Forecast for period t
Ft 1 Forecast for the previous period
= Smoothing constant
At 1 Actual demand or sales from the previous period
3-29
#profevad2015 29
Exponential Smoothing
#profevad2015
Example 3 - Exponential Smoothing
#profevad2015
Picking a Smoothing Constant
Actu
50
al
.4
45 .1
Demand
40
35
1 2 3 4 5 6 7 8 9 10 11 12
Period
#profevad2015
Other Forecasting Methods - Focus
Focus Forecasting
Some companies use forecasts based on a “best current
performance” basis
Apply several forecasting methods to the last several periods
of historical data
The method with the highest accuracy is used to make the
#profevad2015
33
Other Forecasting Methods - Diffusion
Diffusion Models
Historical data on which to base a forecast are not
available for new products
Predictions are based on rates of product adoption and usage
spread from other established products
Take into account facts such as
Market potential
Word-of-mouth
#profevad2015
34
Techniques for Trend
Linear trend equation
Non-linear trends
#profevad2015
35
Linear Trend
A simple data plot can reveal the existence and nature
of a trend
Linear trend equation
Ft a bt
where
Ft Forecast for period t
a Value of Ft at t 0
b Slope of the line
t Specified number of time periods from t 0
#profevad2015
36
Linear Trend Equation
Ft
Ft = a + bt
0 1 2 3 4 5
Ft = Forecast for period t
t
t = Specified number of time periods
a = Value of Ft at t = 0
b = Slope of the line
#profevad2015
Linear Trend Equation
180
160
140
120
100 Actual
80 Trend
60
Underlying
Demand
a 40
20 180
160
0 140
120
1 2 3 4 5 6 7 8 9 10 11 12 13 10014 15 Actual
80 Trend
60
40
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 b (slope)
#profevad2015
Linear Trend Equation
Easy to use: a built-in function
in spreadsheet software
Based on Least Squares method Yt Deviation
used in linear regression, which 180
160
40
periods
Both a and b must be At
recalculated on regular basis to
include new data.
#profevad2015
Estimating slope and intercept
Slope and intercept can be estimated from historical
data
n ty t y
b
n t t
2
2
a
y b t
or y bt
n
where
n Number of periods
y Value of the time series
#profevad2015 3-40
40
Linear Trend Equation Example
Period (t) Sales (Yt) t*Yt t^2
1 67 67 1 Exercise: Calculate b and a by
2 74 148 4 using the sums on the left.
3 102 306 9
4 87 348 16
n tYt t Yt
5 106 530 25
b
n t t
6 86 516 36 2 2
7 117 819 49
8 113 904 64
10 5968 55 998
9 130 1170 81
5.81
10 385 55
10 116 1160 100 2
55 998 5968 385
2 a
Y b t
t
t Yt tYt t n
998 5.81 55
67.84
10
#profevad2015
Linear Trend Equation Example
t y
2
Week t Sales ty
1 1 150 150
2 4 157 314
3 9 162 486
4 16 166 664
5 25 177 885
2
t = 15 t = 55 y = 812 ty = 2499
2
(t) = 225
#profevad2015
Linear Trend Calculation
5 (2499) - 15(812) 12495 -12180
b = = = 6.3
5(55) - 225 275 - 225
812 - 6.3(15)
a = = 143.5
5
y = 143.5 + 6.3t
#profevad2015
Techniques for Seasonality
Seasonality – regularly repeating movements in series
values that can be tied to recurring events
Expressed in terms of the amount that actual values deviate
from the average value of a series
Models of seasonality
Additive
Seasonality is expressed as a quantity that gets added to or
subtracted from the time-series average in order to incorporate
seasonality
Multiplicative
Seasonality is expressed as a percentage of the average (or trend)
amount which is then used to multiply the value of a series in order
to incorporate seasonality
#profevad2015
44
Models of Seasonality
#profevad2015
45
Seasonality
Volume
120
100
80
60
40
20
0
Volume
#profevad2015
Deseasonalized Data
Deseasonalize
90.00
80.00
70.00
60.00
50.00
40.00
30.00
20.00
10.00
0.00
Deseasonalize
#profevad2015
Seasonal Relatives
Seasonal relatives
The seasonal percentage used in the multiplicative seasonally
adjusted forecasting model
Using seasonal relatives
To deseasonalize data
Done in order to get a clearer picture of the nonseasonal (e.g.,
trend) components of the data series
Divide each data point by its seasonal relative
#profevad2015 48
Techniques for Cycles
Cycles are similar to seasonal variations but are of longer
duration
Explanatory approach
Search for another variable that relates to, and leads, the variable
of interest
Housing starts precede demand for products and services
directly related to construction of new homes
If a high correlation can be established with a leading variable,
an equation can be developed that describes the relationship,
enabling forecasts to be made
#profevad2015
49
Associative Forecasting Techniques
Associative techniques are based on the
development of an equation that summarizes
the effects of predictor variables
Predictor variables - variables that can be used to
predict values of the variable of interest
Home values may be related to such factors as home and
property size, location, number of bedrooms, and number of
bathrooms
#profevad2015
50
Simple Linear Regression
Regression - a technique for fitting a line to a set of
data points
Simple linear regression - the simplest form of
regression that involves a linear relationship between
two variables
The object of simple linear regression is to obtain an
equation of a straight line that minimizes the sum of
squared vertical deviations from the line (i.e., the least
squares criterion)
#profevad2015
51
Least Squares Line
yc a bx
where
yc Predicted (dependent) variable
x Predictor (independe nt) variable
b Slope of the line
a Value of yc when x 0 (i.e., the height of the line at the y intercept)
and
n xy x y
b
n x2 x
2
a
y b x or y b x
n
where
n Number of paired observatio ns
#profevad2015 52
Standard Error
Standard error of estimate
A measure of the scatter of points around a regression
line
If the standard error is relatively small, the predictions
using the linear equation will tend to be more accurate
than if the standard error is larger
y y
2
c
Se
n2
where
S e standard error of estimate
y y value of each data point
n number of data points
#profevad2015 53
Correlation Coefficient
Correlation, r
A measure of the strength and direction of relationship between
two variables
Ranges between -1.00 and +1.00
n xy x y
r
n x2 x
2
n y2 y
2
#profevad2015 54
Simple Linear Regression Assumptions
1. Variations around the line are random
2. Deviations around the average value (the line)
should be normally distributed
3. Predictions are made only within the range of
observed values
#profevad2015
55
Forecast Accuracy and Control
Forecasters want to minimize forecast errors
It is nearly impossible to correctly forecast real-world
variable values on a regular basis
So, it is important to provide an indication of the extent
to which the forecast might deviate from the value of the
variable that actually occurs
Forecast accuracy should be an important forecasting
technique selection criterion
Error = Actual – Forecast
If errors fall beyond acceptable bounds, corrective
action may be necessary
#profevad2015
56
Forecast Accuracy Metrics
MAD
Actual t Forecast t
MAD weights all errors
n evenly
Actual t Forecast t
2
MSE weights errors according
MSE to their squared values
n 1
Actual t Forecast t
Actual t
100
MAPE weights errors
MAPE
n according to relative error
#profevad2015 57
Forecast Error Calculation
Actua Forecas
l t (A-F)
Period [|
Error
(A) (F) |Error| Error 2
Error|/Actual]x100
1 107 110 -3 3 9 2.80%
2 125 121 4 4 16 3.20%
3 115 112 3 3 9 2.61%
4 118 120 -2 2 4 1.69%
5 108 109 1 1 1 0.93%
Sum 13 39 11.23%
n=5 n-1 = 4 n=5
MAD MSE MAPE
= 2.6 = 9.75 = 2.25%
#profevad2015 58
Issues to consider:
Always plot the line to verify that a linear
relationship is appropriate
The data may be time-dependent.
If they are
use analysis of time series
use time as an independent variable in a multiple
regression analysis
A small correlation may indicate that other
variables are important
#profevad2015
59
Monitoring the Forecast
Tracking forecast errors and analyzing them can provide useful
insight into whether forecasts are performing satisfactorily
Sources of forecast errors
The model may be inadequate
Irregular variations may have occurred
The forecasting technique has been incorrectly applied
Random variation
Control charts are useful for identifying the presence of non-
random error in forecasts
Tracking signals can be used to detect forecast bias
#profevad2015
60
Choosing a Forecasting Technique
Factors to consider
Cost
Accuracy
Availability of historical data
Availability of forecasting software
Time needed to gather and analyze data and prepare a
forecast
Forecast horizon
#profevad2015
61
Using Forecast Information
Reactive approach
View forecasts as probable future demand
React to meet that demand
Proactive approach
Seeks to actively influence demand
Advertising
Pricing
Product/service modifications
Generally requires either an explanatory model or a subjective
assessment of the influence on demand
#profevad2015
62
Operations Strategy
The better forecasts are, the more able organizations will
be to take advantage of future opportunities and reduce
potential risks
A worthwhile strategy is to work to improve short-term forecasts
Accurate up-to-date information can have a significant effect on
forecast accuracy:
Prices
Demand
#profevad2015 63