Important Announcement

• Quiz 1: 28 January 2025

• Coverage: Introduction to Operations
Management, Competitiveness + Strategy +
Productivity and Forecasting
• Total items – 40 (Concept-based and Problem
solving questions)
Forecasting
DSIOPMA
Decision Sciences and Innovation Department

“The most important things in life are the connections you make with others.”
Learning
Roadmap
Importance Features of
of forecast forecast

Elements of Forecasting
good forecast process

Forecast Forecasting
errors techniques
Why forecasting • Forecasting seeks to find the right balance between
supply and demand
matter?
Why Every Business Needs it?
• Forecasting reduces the risk of uncertainty
Forecast Defined &
its Nature
• Forecast is a statement about
the future value of a variable of
interest
• Forecasts are made with
reference to a specific time
horizon.
• Forecasts affect decisions and
activities throughout an
organization.
 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
 Determine the Establish a time Obtain, clean, and
Steps in the purpose of the
forecast
horizon analyze appropriate
data
Forecasting
Process

Select a forecasting Make the forecast Monitor the

technique forecast errors
 Allowances should be made
for forecast errors
Forecast
Accuracy and
Control Forecast errors should be
monitored
Forecast Accuracy Metrics

MAD =
 Actual t − Forecastt MAD weights all errors evenly

 (Actual t − Forecastt ) MSE weights errors according to

2

MSE = their squared values

n −1

Actual t − Forecastt
 Actual t
100 MAPE weights errors according to
MAPE = relative error
n
Forecast Error Calculation Example
 Solution

Period A F E IErrorI Error2 E/A

1 107 110 -3 3 9 2.8% MAD 2.6
2 125 121 4 4 16 3.2% MSE 9.75
3 115 112 3 3 9 2.6% MAPE 2.26%
4 118 120 -2 2 4 1.7%
5 108 109 -1 1 1 0.9%
13 39 11.3%
Practice

Find:

a) MAD
b) MSE
c) MAPE
Solution
Solution
Forecasting Approaches

• Qualitative forecasting
• permit the inclusion of soft
information
• Quantitative forecasting
• Rely on hard data
• Involves either the projection of
historical data or the
development of associative
methods
 Forecast Techniques

Judgmental Time-series Associative

Qualitative Quantitative
Qualitative Forecast Techniques

Executive opinions

Judgmental
Sales force opinions

Consumer
Basic Forecasting
Techniques:
Qualitative
Methods
https://hbr.org/1971/07/how-to-choose-
the-right-forecasting-technique
 Quantitative Forecast Techniques
Naive

Associative
Time-series

Moving average
Regression
Exponential smoothing

Linear Trend
Basic Forecasting
Techniques:
Time Series and
Projection
https://hbr.org/1971/07/how-to-choose-
the-right-forecasting-technique
Time-Series Forecasts • Time-series forecast – 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

• Analysis of time-series
data involves identifying
the underlying behavior of
the series
 Time-Series Behaviors

Trend

Seasonality

Cycles

Irregular variations

Random variation
Trends and Seasonality

Trend Seasonality
A long-term upward or downward Short-term, fairly regular variations related
movement in data to the calendar or time of day
Cycle and
Variation
• A long-term
upward or
Cycle downward
movement in data

• Due to unusual
Irregular circumstances that
do not reflect
variation typical behavior
Time-Series Forecasting
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 time
period’s value

• Can be used with

• A stable time series
• Seasonal variations
• Trend
Time-Series Forecasting
Naïve Forecast
• Example
• Stable
If August sales of furniture is 40, then September
sales would be 40.

• Seasonal
The forecast for demand for school bags this
school year is equal to demand for school bag last
school year.
Time-Series Forecasting
Naïve Forecast
• Example

• Trend
Suppose the last two values were 50 and 55. What is the forecast for
the next period?

1
2 50
3 55 +5 55+5 = 60

the forecast is equal to the last value of the series plus or minus the difference between the last two values of the series.
 • Averaging techniques smooth variations in
the data
• They can handle step changes or gradual
changes in the level of a series
Time-Series Forecasting • When is this technique appropriate?
Averaging • Techniques
1. Moving average
2. Weighted moving average
3. Exponential smoothing
 • Technique that averages a number of the most recent
actual values in generating a forecast
Moving Average

σ𝑛𝑖=1 𝐴𝑡−𝑖 𝐴𝑡−𝑛 +. . . +𝐴𝑡−2 + 𝐴𝑡−1

𝐹𝑡 = MA𝑛 = =
Time-Series Forecasting

𝑛 𝑛
where
𝐹𝑡 = Forecast for time period 𝑡
MA𝑛 = 𝑛 period moving average
𝐴𝑡−𝑖 = Actual value in period 𝑡 − 𝑖
𝑛 = Number of periods in the moving average
Moving Average Example
Donna’s Garden Supply wants a 3-month moving-average forecast, including a
forecast for next January, for shed sales.

MONTH ACTUAL SHED SALES 3-MONTH MOVING AVERAGE

January 10 10
February 12 12
March 13 13
April 16 (10 + 12 + 13)/3 = 11.67
May 19 (12 + 13 + 16)/3 = 13.67
June 23 (13 + 16 + 19)/3 = 16
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 (29 + 30 + 28)/3 = 28
November 16 (30 + 28 + 18)/3 = 25.33
December 14 (28 + 18 + 16)/3 = 20.67
Moving Averages
Figure X: Data Analysis Dialog Box

© 2019 Cengage. All Rights Reserved.

Moving Averages
Figure Y: Moving Average Dialog Box

© 2019 Cengage. All Rights Reserved.

Moving Averages

Figure Z: Excel Output for

Moving Average Forecast
for Gasoline Data

© 2019 Cengage. All Rights Reserved.

 Practice:
Compute MAD for periods 4 through 6 using 3-month moving average.

3-month
Month Demand Moving Average
1 650
2 700
3 810
4 800
5 900
6 700

© 2019 Cengage. All Rights Reserved.

 Solution:
3-month Moving Absolute
Demand Average Deviation Deviation

800 720 800-720 = 80 80

900 770 900-770 = 130 130

700 837 700-837 = -137 137

Sum of Absolute Deviation = 80+130+137 = 347

MAE = 347/3 = 115.7

© 2019 Cengage. All Rights Reserved.

 Important Notes

► Increasing n smooths the forecast but makes it

less sensitive to changes
► Does not forecast trends well
► Requires extensive historical data
 Create a time series plot
(scatter diagram) and make a
Practice 1 linear trend forecast for periods
11-13 using the bicycle sales
data.

© 2019 Cengage. All Rights Reserved.

Practice 2:

Freight car loadings over a 12-year period at a busy port are as follows:

a. Determine a linear trend line for expected freight car loadings.

b. Use the trend equation to predict expected loadings for weeks 19 and 20.
c. The manager intends to install new equipment when the volume exceeds 800 loadings per
week. Assuming the current trend continues, the loading volume will reach that level in
approximately what week?
 The manager of a seafood restaurant was asked to establish
a pricing policy on lobster dinners. Experimenting the prices
produced the flowing data:

Practice 3

a. Plot the data and the regression line on the same graph.
b. Determine the correlation coefficient and interpret it.
Round r to four decimals.

© 2019 Cengage. All Rights Reserved.

 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 )
Time-Series Forecasting

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.
 Sample Problem
Given the following demand data:
a. Determine the demand for June using a weighted moving
average with a weight of .40 for the most recent period, .30 for
the next most recent, .20 for the next and .10 for the next.
b. If the actual demand for June is 49, forecast demand for July
using the same weights as in part a.

Period Demand Weights Forecast Period Demand Weights Forecast

Jan 53 Jan 53
Feb 50 0.1 5 Feb 50
March 53 0.2 10.6 March 53 0.1 5.3
April 50 0.3 15 April 50 0.2 10
May 51 0.4 20.4 May 51 0.3 15.3
June 51 June 49 0.4 19.6
July July 50.2
 Exponential Smoothing • A weighted averaging method that is based on the previous
forecast plus a percentage of the forecast error

Ft = Ft −1 +  ( At −1 − Ft −1 )
Time-Series Forecasting

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
Exponential
Smoothing Example
In January, a local car
dealer predicted February
demand for 142 Ford
Mustangs. Actual February
demand was 153 autos.
Using a smoothing constant
chosen by management of a
= .20, the dealer wants to
forecast March demand
using the exponential
smoothing model.
 Exponential Smoothing Example

Predicted demand = 142 Ford Mustangs

Actual demand = 153
Smoothing constant a = .20

New forecast = 142 + .2(153 – 142)

 Exponential Smoothing Example

Predicted demand = 142 Ford Mustangs

Actual demand = 153
Smoothing constant a = .20

New forecast = 142 + .2(153 – 142)

= 142 + 2.2
= 144.2 ≈ 144 cars
 Exponential Smoothing Practice

Suppose the previous forecast was 45 units, actual demand

was 42 units, and  .10, what is the forecast value?

Suppose the actual demand turns out to be 43, what would

be the next forecast?
 • A simple data plot can reveal the existence and nature of a
trend
• Linear trend equation
Linear Trend
Time-Series Forecasting

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
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
Trend Line Equation
Example Week Unit Sales
1 700
Computer sales for a
2 724
Hong Kong-based firm
over the last 10 weeks 3 720
are shown. Plot the
4 728
data, and visually
check whether a linear 5 740
trend line would be
6 742
appropriate. Then
determine the equation 7 758
of the trend line, and 8 750
forecast sales for
weeks 11 and 12. 9 770
10 775
Trend Line
Equation Example

x = time
y = sales
 Trend Line Equation Example

= 10(41,358) – 55(7,407) =
6,195 = 7. 51
10(385) - 55(55) 825

= 7,407- 7.51(55)
= 699. 40
10
The trend line is Ft = 699.40 + 7.51t, where t = 0 for period 0
 Trend Line Equation Example

Substituting the values of t into equation, the forecasts for the next two
periods (t = 11 & t =12)

F11 = 699.40 + 7.51(11) = 782.01

F12 = 699.40 + 7.51(12) = 789.52

 Trend Line
Equation Example
 Data Analysis - Regression

SUMMARY OUTPUT

Regression Statistics
Multiple
R 0.969606 Week Unit Sales
R Square 0.940135 1 700
Adjusted
R Square 0.932652 2 724
Standard 3 720
Error 6.085004
Observati
4 728
ons 10 5 740
6 742
ANOVA
Significan 7 758
df SS MS F ce F 8 750
Regressio
n 1 4651.882 4651.882 125.6339 3.6E-06 9 770
Residual 8 296.2182 37.02727 10 775
Total 9 4948.1 11 782.00 699.4+(7.51*11)
Coefficien Standard Lower Upper Lower Upper 12 789.51 699.4+(7.51*12)
ts Error t Stat P-value 95% 95% 95.0% 95.0%
Intercept 699.4 4.156849 168.2524 1.74E-15 689.8143 708.9857 689.8143 708.9857
Sales 7.509091 0.669937 11.20865 3.6E-06 5.964214 9.053968 5.964214 9.053968
 800

790

780

770

760

750
Sales

740

730

720

710

700

690
0 2 4 6 8 10 12 14
Time
 • Are based on the development of an equation
Associative that summarizes the effects of predictor
variables
Forecasting • Predictor variables - variables that can be
Techniques used to predict values of the variable of
interest
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 variable
• Aims 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)
 yc = a + bx
where
Least Squares Line 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  x 2 − ( x )
2

a=
 y − b x
or y − b x
n
where
n = Number of paired observations
Simple Linear Regression
Assumptions

• Variations around the line are random

• Deviations around the average value (the
line) should be normally distributed
• Predictions are made only within the
range of observed values
Simple Linear Regression Wellness
MOS Burgers
in Baguio
chain
Hamburgers
has a chainhas
city. Sales
of stores figures and
in Baguio
a
of stores
city.
profit for the stores are given in
Example 1 Sales figures
the table.
equation
Obtainanda profit
regression for the
stores arefor giventhe data
in theand predict
profit for a store assuming sales of
following table. Obtain a
Php 10 million.
regression equation for the
Units Sales Profits
data(inand predict
Php millions) profit
(in Php for a
millions)
7
store assuming 0.15of Php 10
sales
2 0.10
million.6 0.13
4 0.15
14 0.25
15 0.27
16 0.24
12 0.20
14 0.27
20 0.44
15 0.34
7 0.17
Simple Linear Regression Solution
Wellness Hamburgers has a chain of stores in
Baguio city. Sales figures and profit for the stores
are given in the following table. Obtain a
regression equation for the data and predict
Linear model profit for a store assuming sales of Php 10
million.
 Simple Linear Regression Solution
(Using Data Analysis)
Wellness Hamburgers has a chain of stores in Baguio city.
Sales figures and profit for the stores are given in the
following table. Obtain a regression equation for the data
and predict profit for a store assuming sales of Php 10
million.
Simple Linear Regression Solution
Wellness Hamburgers has a chain of stores in Baguio city.
Sales figures and profit for the stores are given in the
following table. Obtain a regression equation for the data
and predict profit for a store assuming sales of Php 10
million.
Correlation Coefficient
• Correlation, r
• A measure of the strength and direction of relationship between two variables

n( xy ) − ( x )( y )
r=
( )
n  x 2 − ( x )
2
( )
n  y 2 − ( y )
2

• r2, square of the correlation coefficient, coefficient of determination

• A measure of the percentage of variability in the values of y that is “explained” by
the independent variable
 • We can model a time series with a seasonal pattern by
treating the season as a dummy variable.
Seasonality without Trend • Illustration:
• Consider the data on the number of umbrellas
sold
© 2019 Cengage. All Rights Reserved.
 Year Quarter Sales
1 1 125
Illustration (cont.): 2 153
3 106
• k-1 dummy variables 4 88
are required to model 2 1 118
a categorical variable 2 161
that has k levels. 3 133

Umbrella Sales
4 102
3 1 138
• Thus, to model the 2 144
seasonal effects in the 3 113
4 80
umbrella time series
4 1 109
we need 4-1 = 3
2 137
dummy variables 3 125
4 109
5 1 130
2 165
3 128
4 96
© 2019 Cengage. All Rights Reserved.
 Umbrella Sales

© 2019 Cengage. All Rights Reserved.

 Seasonality without Trend

• The three dummy variables can be coded as follows:

Qtr1𝑡 = 1 if period 𝑡 is quarter 1; 0 otherwise.

Qtr2 𝑡 = 1 if period 𝑡 is quarter 2; 0 otherwise.
Qtr3𝑡 = 1 if period 𝑡 is quarter 3; 0 otherwise.

• General form of the equation relating the number of umbrellas sold

to the quarter the sales take place:

© 2019 Cengage. All Rights Reserved.

Table 8.11: Umbrella Sales Time Series with Dummy Variables

Period Year Quarter Qtr1 Qtr2 Qtr3 Sales

1 1 1 1 0 0 125
2 2 0 1 0 153
3 3 0 0 1 106
4 4 0 0 0 88
5 2 1 1 0 0 118
6 2 0 1 0 161
7 3 0 0 1 133
8 4 0 0 0 102

© 2019 Cengage. All Rights Reserved.

 Table 8.11: Umbrella Sales Time Series with Dummy Variables (cont.)
Period Year Quarter Qtr1 Qtr2 Qtr3 Sales
9 3 1 1 0 0 138
10 2 0 1 0 144
11 3 0 0 1 113
12 4 0 0 0 80
13 4 1 1 0 0 109
14 2 0 1 0 137
15 3 0 0 1 125
16 4 0 0 0 109

© 2019 Cengage. All Rights Reserved.

Table 8.11: Umbrella Sales Time Series with Dummy Variables (cont.)

Period Year Quarter Qtr1 Qtr2 Qtr3 Sales

17 5 1 1 0 0 130
18 2 0 1 0 165
19 3 0 0 1 128
20 4 0 0 0 96

© 2019 Cengage. All Rights Reserved.

 𝑦ො𝑡 = 95.0 = 29.0Qtr1t +57.0Qtr2t + 26.0Qtr3t

© 2019 Cengage. All Rights Reserved.

 Seasonality without Trend:

𝑦ො𝑡 = 95.0 = 29.0Qtr1t +57.0Qtr2t + 26.0Qtr3t

Quarter 1 Sales = 95.0 + 29(1) + 57.0(0) + 26.0(0) = 124

Quarter 2 Sales = 95.0 + 29(0) + 57.0(1) + 26.0(0) = 152
Quarter 3 Sales = 95.0 + 29(0) + 57.0(0) + 26.0(1) = 124
Quarter 4 Sales = 95.0 + 29(0) + 57.0(0) + 26.0(0) = 95

© 2019 Cengage. All Rights Reserved.

Final Word

© 2019 Cengage. All Rights Reserved.