FORECASTING

Management science by Anderson

 FORECASTING
Objectives:
1. List the types of the major forecasting approaches
and their characteristics;
2. Describe the need for forecasting and the role it
plays in decision making;
3. Enumerate the two Classifications of forecasting;
4. Solve problems using exponential smoothing and
moving average;
5. Determine when a forecast can be improved.

Management science by Anderson

 FORECASTING
FORECASTING
- is the method of calculating or predicting
some future events or conditions usually as a
result of some rational study and analysis of
available prediction of what will happen in
the future.
- a method for translating past experience into
estimates of the future.
- A process of estimating the unknown.
A Forecast is a prediction, estimate, or
determination of what will occur in the
future based on a certain set of factors.
sales, production
Interest rates
Funds
GNP
Technological status
Inventory
Scheduling
Facility layout
Workforce, and so on
 Suppose you are asked to provide quarterly
forecasts of sales for one of your company’s products
over the one – year period. Various factors like
• production schedules
• raw materials purchasing, policies, and
• quotas
will all be affected by the quarterly forecasts you
provide.
A good forecast may result in quality planning
and increased profit, whereas, a poor forecast may
result in poor outcome and may decrease
profit/increased cost.
Classification of Forecasting Methods

Forecasting Methods

Quantitative Methods Qualitative Methods

Time Series Causal Delphi

Method

Moving Regression Jury of

Averages Analysis Executive
Opinion
Exponential Multiple
Smoothing Regressions Sales Force
Composite
Trend Consumer
Projection Market
Quantitative Approaches in Decision Making by Arao, Rosalina et. al Survey
https://www.slideshare.net/anandsubramaniam/forecasting-techniques-2663882
 Qualitative and Quantitative Forecasting
Forecasting methods can be classified as qualitative or
quantitative:
1. Qualitative methods – involved the use of expert
judgment. Such methods are appropriate when
historical data on the variable being forecast are
either unavailable or not applicable.
2. Quantitative methods – can be used when,
i. Past information about the variable being forecast
is available.
ii. The information can be quantified.
iii. It is reasonable to assume that past is prologue.
i.e., the pattern of the past will continue into the
future).
Factors a forecast may be based:
past data
Opinion or Judgment
Company data or
Perceived pattern related to time
 Time Series
A time series is a sequence of observations on a
variable measured at successive points in time or
over successive periods of time.
The measurements may be taken every hour, day,
week, month, or year. Or at any other regular
interval.
The pattern of the data is an important factor in
understanding how the time series has behaved in
the past.
If such behavior can be expected to continue in the
future we can use it to guide us in selecting an
appropriate forecasting method.
 Triple A Bond interest rates Enrollment in Thousands for
for 12 months a State University
Month Interest Year Enrollment
Rates
1 6.5
1 9.5 2 8.1
2 9.3 3 8.9
3 9.4 4 10.8
4 9.6 5 11.7
5 9.8 6 12.3
6 9.7 7 13.9
7 9.8 8 18.1
8 10.5 9 19.00
9 9.9
10 9.7
11 9.6
12 9.6

1. Construct a time series plot.

2. What type of pattern exists in the data?
3. What is the quarterly forecast of the data for the next n periods?
https://www.google.com.ph/search
 Time Series Plot
A time series plot is a graphical
presentation of the relationship between time
and the time series variable.
time – is represented on the horizontal axis
values of the time series variable (sales,
demand, revenue, etc.) - are represented on
the vertical axis
https://www.google.com.ph/search
Exploring Time Series Data Patterns:
There are typically four general types
of patterns: horizontal, trend, seasonal, and
cyclical.

Horizontal Trend Seasonal Cyclical

Pattern Pattern Pattern Pattern
 Horizontal Pattern
A horizontal pattern exists when the data
fluctuate randomly around a constant mean
over time.
Sales (1000s
To illustrate, consider the following Week of gallons)
1 17
data: 2 21
The table on the right shows 3 19
the number of gallons of gasoline 4 23
5 18
(in 1000s) sold by a gasoline 6 16
distributor over the past 12 weeks. 7 20
The average value or mean for this 8 18
9 22
time series is 19.25 or 19, 250
10 20
gallons per week. 11 15
12 22
 Week Sales (1000s of gallons)
Horizontal Pattern 1 17
The figure below 2 21
3 19
shows a time series plot 4 23
for these data. Note how 5 18
6 16
the data fluctuate around 7 20
the sample mean of 19, 8 18
9 22
250 gallons. Although 10 20
random variability is 11 15
12 22
present we would say
Sales (1000s of gallons)
that these data follow a Sales (1000s of gallons) 25

horizontal pattern. 20

15

10 Sales (1000s of
gallons)
5

0
1 2 3 4 5 6 7 8 9 10 11
Week
 Stationary Time Series
The term stationary time series are used to
denote a time series whose statistical
properties are independent of time. In
particular, this means that
1. The process generating the data has a
constant mean.
2. The variability of the time series is
constant over time.
 Changes in business conditions often result
in a time series with a horizontal pattern that
shifts to a new level at some point in time.
For instance, suppose the gasoline
distributor signs a new contact to provide
gasoline in week 13. With this new contract, the
distributor naturally expects to see a substantial
increase in weekly sale starting in week 13.
Week Sales (1000s of gallons) Week Sales (1000s of gallons)
1 17 12 22
2 21 13 31
3 19 14 34
4 23 15 31
5 18 16 33
6 16 17 28
7 20 18 32
8 18 19 30
9 22 20 29
10 20 21 34
11 15 22 33
Gasoline Sales Time Series Plot After Obtaining
the Contract with the Vermont state Police
40
35
Sales (1000s of gallons)

30
25
20
15 Sales (1000s of gallons)
10
5
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Week
Trend Pattern
A trend pattern exists when the series
show a gradual shifts or movements to relatively
higher or o lower values over a longer period of
time. It is usually the result of long-term factors
such as populations increases or decreases,
shifting demographic characteristics of the
population, improving technology, and /or
changes in consumer preferences.
 Sales
Year (1000s )
1 21.6 Bicycle Sales Time Series
2 22.9 35

3 25.5 30

25
4 21.9
Sales (1000s )
20
5 23.9 15 Sales (1000s )
6 27.5 10

7 31.5 5

8 29.7 0
1 2 3 4 5 6 7 8 9 10
9 28.6 Year

10 31.4
Year Revenue
1 23.1 Cholesterol Drug Rrevenue Time Series Plot ($ Millions)
120
2 21.3
3 27.4 100

4 34.6 80

5 33.8
Revnue

60
6 43.2 Revenue

7 59.5 40

8 64.4 20

9 74.2 0
10 99.3 1 2 3 4 5
Year
6 7 8 9 10
Seasonal Pattern
A trend pattern where movements in
historical data are analyzed over multiple years.
Seasonal patterns are recognized by observing
recurring patterns over successive period of
time.
UMBRELLAS SALES TIME SERIES
Year Quarter Sales Year Quarter Sales
1 1 125 4 1 109
2 153 2 137
3 106 3 125
4 88 4 109
2 1 118 5 1 130
2 161 2 165
3 133 3 128
4 102 4 96
3 1 138
2 144
3 113
4 80

UMBRELLAS SALES TIME SERIES

180
160
140
120
100
Sales

80
Sales
60
40
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Time Period
Trend and Seasonal Pattern
This pattern includes both the trend and
seasonal pattern.
QUARTERLY TELEVISION SET SALES TIME SERIES
Year Quarter Sales (1000s) Year Quarter Sales (1000s)
1 1 4.8 3 1 6
2 4.1 2 5.6
3 6 3 7.5
4 6.5 4 7.8
2 1 5.8 4 1 6.3
2 5.2 2 5.9
3 6.8 3 8
4 7.4 4 8.4
QUARTERLY TELEVISION SET SALES TIME SERIES PLOT
10
QUARTERLY TELEVISION SET

8
6
SALES (1000s)

4
Sales (1000s)
2
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Period
 QUARTERLY TELEVISION SET SALES TIME SERIES
Year Quarter Sales (1000s) Year Quarter Sales (1000s)
1 1 4.8 3 1 6
2 4.1 2 5.6
3 6 3 7.5
4 6.5 4 7.8
2 1 5.8 4 1 6.3
2 5.2 2 5.9
3 6.8 3 8
4 7.4 4 8.4

The table above and the QUARTERLY TELEVISION SET SALES TIME SERIES PLOT
9
corresponding time series on the left

QUARTERLY TELEVISION SET SALES (1000s)

8
show quarterly television set sales for a 7
particular manufacturer over the past 6

four years. It can be observed that an 5

4
increasing trend is present. The figure Sales (1000s)
3
also indicates that sales are lowest in 2
the second quarter of each year and 1

highest in quarters 3 and 4. 0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Thus, we conclude that a Period
trend pattern and seasonal pattern
exists for television sales.
Cyclical Pattern
This pattern exists if the time series plot
shows an alternating sequence of points below
and above the trend line that lasts for more than
one year.
For example, periods of moderate inflation
followed by periods of rapid inflation can lead to
a time series that alternates below and above a
generally increasing trend line.
Week Sales (1000s ) Time Series of Housing Cost (1000s )
1 7 35

2 15.7 30

3 16.7 25

4 25

Sales (1000s )
20

5 18 15 Sales (1000s )

6 23 10

7 10 5

8 29.7 0
1 2 3 4 5 6 7 8 9 10
9 26 Year

10 17
 FORECAST ACCURACY
Forecast accuracy is used to determine how well
a particular forecasting method is able to
reproduce the time series data that are
already available.
By selecting the method that is most accurate
for the data already known, we hope to
increase the likelihood that we will obtain
more accurate forecasts for the future time
periods.
Two Forecasting Methods:
1. Naïve Forecasting Method
An approach that uses the most recent data
value as the forecast for the next period. It is the
simplest method of forecasting.
2. Average of Past Values (Average of Historical
Values)
An approach that uses the average of all the
historical data available as the forecast for the next
period.
Consider the data below:
Period Demand
1 47
2 52
3 67
4 77
5 72

Naïve Average
Period Demand Forecasting Forecasting
1 47
2 52 47 47.00
3 67 52 49.50
4 77 67 55.33
5 72 77 60.75
 Measures of Forecast Accuracy
1. Mean or Average of the Forecast Errors (MFE)
2. Mean Absolute Error (MAE)
3. Mean square Error (MSE)
4. Mean Absolute Percentage Error (MAPE)
n

∑
n n
 et 
∑
2
e
et t ∑ 
Y
 100
MAE = t = k +1 MSE = t= k +1 t= k +1  t 
n − k MAPE =
n − k n − k
 FORECAST ERROR et
Forecast Error et - is the difference between the
actual and forecasted values for period t. It is
the key concept associated with measuring
forecast accuracy.
e t = Y t − Ft
where :
t = period
e t = error at period t
Y t = actual forecast
Ft = forecasted value
 Measures of Forecast Accuracy
Mean or Average of the Forecast Errors (MFE)
n

∑e
t = k +1
t
MFE =
n−k

where :
t = period
e t = error at period t
k = number of periods at the beginning of the time series
for which we cannot produce a naive forecast.
n = total number of periods
Mean Absolute Error (MAE)
n

∑
t = k +1
et
MAE =
n−k

where :
t = period
e t = absolute value of error at period t
k = number of periods at the beginning of the time series
for which we cannot produce a naive forecast.
n = total number of periods
Mean square Error (MSE)
n

∑
t = k +1
e 2t
MSE =
n−k

where :
t = period
e 2t = squared error at period t
k = number of periods at the beginning of the time series
for which we cannot produce a naive forecast.
n = total number of periods
Mean Absolute Percentage Error (MAPE)

n
 et 
∑  100
t = k +1  Yt 
MAPE =
n−k
where :
t = period
e t = error at period t
k = number of periods at the beginning of the time series
for which we cannot produce a naive forecast.
n = total number of periods
 FORECAST ACCURACY
To illustrate, consider the following Sales (1000s
data: Week of gallons)
1 17
The table on the right shows 2 21
the number of gallons of gasoline 3 19
(in 1000s) sold by a gasoline 4 23
5 18
distributor over the past 12 weeks. 6 16
Make a forecast about the data 7 20
using the naïve forecasting and the 8 18
average of the past, then compare 9 22
10 20
which of the two methods gives the 11 15
more accurate forecast using MAE, 12 22
MSE, and MAPE.
 Absolute
Absolute Value of
Time Series Forecast Value of Squared Percentage Percentage
Week Value Forecast Error forecast Error Forecast Error Error Error
1 17
2 21 17 4 4 16 19.05 19.05
3 19 21 -2 2 4 -10.53 10.53
4 23 19 4 4 16 17.39 17.39
5 18 23 -5 5 25 -27.78 27.78
6 16 18 -2 2 4 -12.50 12.50
7 20 16 4 4 16 20.00 20.00
8 18 20 -2 2 4 -11.11 11.11
9 22 18 4 4 16 18.18 18.18
10 20 22 -2 2 4 -10.00 10.00
11 15 20 -5 5 25 -33.33 33.33
12 22 15 7 7 49 31.82 31.82
Total 5 41 179 1.19 211.69
n n

∑
t = k +1
et ∑ e 2t
∑
n
 et 
 100
MAE = MSE = t = k +1 t = k +1  Y t 
n−k n−k MAPE =
n − k
41 179 211.69
= = =
11 11 11
= 3.73 = 19.24%
= 16.27
 Absolute Squared
Time Series Forecast Value of Forecast Percentage Absolute Value of
Week Value Forecast Error forecast Error Error Error Percentage Error
1 17
2 21 17.00 4.00 4.00 16.00 19.05 19.05
3 19 19.00 0.00 0.00 0.00 0.00 0.00
4 23 19.00 4.00 4.00 16.00 17.39 17.39
5 18 20.00 -2.00 2.00 4.00 -11.11 11.11
6 16 19.60 -3.60 3.60 12.96 -22.50 22.50
7 20 19.00 1.00 1.00 1.00 5.00 5.00
8 18 19.14 -1.14 1.14 1.31 -6.35 6.35
9 22 19.00 3.00 3.00 9.00 13.64 13.64
10 20 19.33 0.67 0.67 0.44 3.33 3.33
11 15 19.40 -4.40 4.40 19.36 -29.33 29.33
12 22 19.00 3.00 3.00 9.00 13.64 13.64
Total 4.52 26.81 89.07 2.75 141.34
n

∑  et 
n n
∑ e 2t
et
t = k +1
∑   100
t = k +1  Y t 
MAE = t = k +1 MSE = MAPE =
n − k n − k n − k
26.81 89.07 141.34
= = =
11 11 11
= 2.44 = 8.10 = 12.85%
Comparing Naïve Forecasting Method and
Average of the Past Values
Naïve Method Average of the Past Values
MAE 3.73 2.44
MSE 16.27 8.10
MAPE 19.24% 12.85%

Take home Activity(Part 1 of Quiz 6):

1. Make a research about the difference between the Naïve
Forecasting Method and Average of the Past Values.
2. Based on the data and the results of the forecast
accuracy of the data, write your explanation why is it that
in the gasoline time series problem, the average of the
past values is more appropriate compared to the naïve
forecasting. (please indicate your references)
 Other Forecasting Methods or
Time Series Methods
The next set of methods will illustrate three
forecasting methods that are appropriate for a time
series with a horizontal pattern: the moving averages,
weighted moving averages, and exponential smoothing.
These methods are also capable of adapting well
to changes in the level of a horizontal pattern such as
what we saw with the extended gasoline sales time
series.
Because the objectives of these methods is to
“smooth out” random fluctuations in the time series,
they are referred to as smoothing methods.
Other Forecasting Methods or
Time Series Methods
1. Moving Averages(Simple Moving
Averages)
2. Weighted Moving Averages
3. Simple Exponential Smoothing
 Forecasting Methods or Time Series Methods

A Moving Average is the un-weighted average of

a consecutive number of data points. It uses the
average of the most recent k data values in the
time series as the forecast for the next period.
To compute a simple moving average, we simply
choose the number of items in the time series
data to include in the average.

Moving Average =
∑ (most recent k data values)
k
∧
Y t +1 =
∑ (most recent k data values)
k
Y t - k +1 + ... + Y t −1 + Y1
=
k
t

∑Y
i = t − k +1
i
=
k
∧
Y t +1 = Forecast of the time series for period t + 1
Yi = Actual values of the time series in period i
k = number of periods of the time series data
used to generate the forecast
 Moving Averages
To illustrate, consider the Sales (1000s
Week of gallons)
following data: 1 17
The table on the right 2 21
3 19
shows the number of gallons of 4 23
gasoline (in 1000s) sold by a 5 18
6 16
gasoline distributor over the 7 20
past 12 weeks. Compute the 3- 8 18
9 22
week moving averages of the 10 20
time series. 11 15
12 22
Moving Averages
Squared
Time Series Forecast Absolute Value Forecast Percentage Absolute Value of
Week Value Forecast Error of forecast Error Error Error Percentage Error
1 17
2 21
3 19
4 23 19 4.00 4.00 16.00 17.39 17.39
5 18 21 -3.00 3.00 9.00 -16.67 16.67
6 16 20 -4.00 4.00 16.00 -25.00 25.00
7 20 19 1.00 1.00 1.00 5.00 5.00
8 18 18 0.00 0.00 0.00 0.00 0.00
9 22 18 4.00 4.00 16.00 18.18 18.18
10 20 20 0.00 0.00 0.00 0.00 0.00
11 15 20 -5.00 5.00 25.00 -33.33 33.33
12 22 19 3.00 3.00 9.00 13.64 13.64
Total 0.00 24.00 92.00 -20.79 129.21
n n
∑ e
∑ e 2 n
 e t 
∑
t
t = k +1 t   100
MAE = MSE = t = k + 1
t = k + 1  Y t 
n − k n − k MAPE =
24 n − k
= 92
9 = 129.21
9 =
= 2 . 67 9
= 10.22 = 14.36%
 Forecasting Methods or Time Series Methods
A Weighted Moving Averages is a moving
average that involves selecting a different
weight for each data value in the moving
average and then computing a weighted
average of the most recent k values as the
forecast.
∧
Y t +1 = w t Y t + w t −1 Y t − 1 + ... + w t − k +1 Y t − k +1
∧
Y t +1 = Forecast of the time series for period t + 1
Yt = Actual values of the time series in period t
w t = weight applied to the actual time series value for period t
k = number of periods of the time series data
used to generate the forecast
 Weighted Moving Averages
To illustrate, consider the Sales (1000s
Week of gallons)
following data: 1 17
The table on the right 2 21
3 19
shows the number of gallons of 4 23
gasoline (in 1000s) sold by a 5 18
6 16
gasoline distributor over the 7 20
past 12 weeks. Determine the 8 18
9 22
weighted moving average using 10 20
wt = 1/6, wt-1 = 2/6, wt-2 = 3/6. 11 15
12 22
. Note:
It should be noted that the most recent observation receives the
largest weight and the weight decreases with the relative age of
the data values.
Weighted Moving Averages using wt = 1/6, wt-1 = 2/6, wt-2 = 3/6.
Absolute
Time Value of Squared Absolute Value
Series Forecast forecast Forecast Percentage of Percentage
Week Value Forecast Error Error Error Error Error
1 17
2 21
3 19
4 23 19.33 3.67 3.67 13.44 15.94 15.94
5 18 21.33 -3.33 3.33 11.11 -18.52 18.52
6 16 19.83 -3.83 3.83 14.69 -23.96 23.96
7 20 17.83 2.17 2.17 4.69 10.83 10.83
8 18 18.33 -0.33 0.33 0.11 -1.85 1.85
9 22 18.33 3.67 3.67 13.44 16.67 16.67
10 20 20.33 -0.33 0.33 0.11 -1.67 1.67
11 15 20.33 -5.33 5.33 28.44 -35.56 35.56
12 22 17.83 4.17 4.17 17.36 18.94 18.94
Total 0.50 26.83 103.42 -19.17 143.93
 Forecasting Methods or Time Series Methods
An Exponential Smoothing also uses a weighted
average of past time series values as a
forecast; it is a special case of the weighted
moving averages method in which we select
only one weight – the weight for the most recent
observation. The weights for the other data
values are computed automatically and become
smaller as the observations move farther into
the past. The exponential smoothing model
follows.
∧ ∧
Y t + 1 = α Y t + (1 − α ) Y t
 ∧ ∧
Y t + 1 = α Y t + (1 − α ) Y t

∧
Y t +1 = Forecast of the time series for period t + 1
Yt = Actual values of the time series in period t
∧
Yt = forecast of the time series for period t
α = smoothing constant (0 ≤ α ≤ 1)
 Exponential Smoothing
To illustrate, consider the Sales (1000s
Week of gallons)
following data: 1 17
The table on the right 2 21
3 19
shows the number of gallons of 4 23
gasoline (in 1000s) sold by a 5 18
6 16
gasoline distributor over the 7 20
past 12 weeks. Find the simple 8 18
9 22
exponential smoothing using 10 20
an α of 0.20. 11 15
12 22
 Time Series Squared
Week Value Forecast Forecast Error Forecast Error
1 17
2 21 17.00 4.00 16.00
3 19 17.80 1.20 1.44
4 23 18.04 4.96 24.60
5 18 19.03 -1.03 1.07
6 16 18.83 -2.83 7.98
7 20 18.26 1.74 3.03
8 18 18.61 -0.61 0.37
9 22 18.49 3.51 12.34
10 20 19.19 0.81 0.66
11 15 19.35 -4.35 18.94
12 22 18.48 3.52 12.38
Total 10.92 98.80
 Trend-Adjusted Exponential Smoothing
A variation of simple exponential smoothing can be
used hen a time series exhibits a linear trend. It is called
trend-adjusted exponential smoothing or, sometimes, double
smoothing, to differentiate it from simple exponential
smoothing, which is appropriate only when data vary around
an average or have step or gradual changes.
If a series exhibits trend, and simple smoothing is used on it,
the forecast will all lag the trend
• if the data are increasing, each forecast will be too low;
• If the data are decreasing, each forecast will be too high.

Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution
without the prior written consent of McGraw-Hill Education
 Trend-Adjusted Exponential Smoothing

• The trend adjusted forecast consists of two

components
– Smoothed error
– Trend factor
TAF t +1 = S t + Tt
where
S t = Previous forecast plus smoothed error
Tt = Current trend estimate

Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution
without the prior written consent of McGraw-Hill Education
 Trend-Adjusted Exponential Smoothing (cont.)

• Alpha and beta are smoothing constants

• Trend-adjusted exponential smoothing has the
ability to respond to changes in trend
TAF t +1 = S t + Tt
S t = TAF t + α (At − TAF t )
Tt = Tt −1 + β (TAF t − TAF t−1 − Tt −1 )

Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution
without the prior written consent of McGraw-Hill Education
Trend-Adjusted Exponential Smoothing

TAF t +1 = S t + Tt
where
S t = Previous forecast plus smoothed error
Tt = Current trend estimate

TAF t +1 = S t + Tt
S t = TAF t + α (At − TAF t )
Tt = Tt −1 + β (TAF t − TAF t−1 − Tt −1 )

Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution
without the prior written consent of McGraw-Hill Education
Process:
1. There must be a historical data.
2. Identify the α and β of the problem.
3. Compute the initial trend estimate and the starting forecast
of the current period.
Initial trend estimate = the average of the net changes from 1 – k
Initial Forecast = At-1 + initial trend estimate
4. Compute for the St, St = TAFt + α(At - TAFt).
5. Compute the next forecast TAFt+1 = St + Tt .
6. Compute for St and Tt. Let t + 1 be the next current period t.
Tt = Tt-1 + β(TAFt – TAFt-1 - Tt-1)
7. Continue the process 5 and 6 until the desired forecast is
computed.
8.
 Example:
Cell phone sales for a Week Unit Sales
California-based firm over the last 10
1 700
weeks are shown in the following
2 724
table. Plot the data, and visually
check to see if a linear trend line 3 720
would be appropriate. If it is, use the 4 728
trend-adjusted exponential 5 740
smoothing to obtain forecasts for 6 742
periods 6 though 11, with α = 0.40 7 758
and β = 0.30. Based on the net 8 750
change from periods 1 t0 4. 9 770
10 775
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution
without the prior written consent of McGraw-Hill Education
For purposes of illustration, the original data, the trend
line, and the two projections (forecasts) are shown on
the following graph:

Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution
without the prior written consent of McGraw-Hill Education
Week Unit Sales TAFt St Tt
1 700
2 724
3 720
4 728
5 740
6 742
7 758
8 750
9 770
10 775

Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution
without the prior written consent of McGraw-Hill Education