4 Forecasting

Outline
 What Is Forecasting?
 Forecasting Time Horizons
 The Strategic Importance of Forecasting
 Forecasting Approaches
Qualitative Methods
Quantitative Methods
 Outline – Continued
 Time-Series Forecasting
 Decomposition of a Time Series
 Naive Approach
 Moving Averages
 Exponential Smoothing
 Exponential Smoothing with Trend Adjustment
 Trend Projections
 Seasonal Variations in Data
 Cyclical Variations in Data
 Outline – Continued
 Associative Forecasting Methods:
Regression and Correlation Analysis
 Monitoring and Controlling Forecasts
 Adaptive Smoothing
 Focus Forecasting
 Forecasting in the Service Sector
Forecasting Provides a Competitive Advantage
for Disney
► Global portfolio includes parks in Hong Kong, Paris,
Tokyo, Orlando, and Anaheim
► Revenues are derived from people – how many visitors
and how they spend their money
► Daily management report contains only the forecast and
actual attendance at each park
► Disney generates daily, weekly, monthly, annual, and 5-
year forecasts
► Forecast used by labor management, maintenance,
operations, finance, and park scheduling
► Forecast used to adjust opening times, rides, shows,
staffing levels, and guests admitted
 Forecasting Provides a Competitive
Advantage for Disney
 20% of customers come from outside the USA
 Economic model includes gross domestic product,
cross-exchange rates, arrivals into the USA
 A staff of 35 analysts and 70 field people survey 1
million park guests, employees, and travel professionals
each year
 Inputs to the forecasting model include airline specials,
Federal Reserve policies, Wall Street trends,
vacation/holiday schedules for 3,000 school districts
around the world
 Average forecast error for the 5-year forecast is 5%
 Average forecast error for annual forecasts is between
0% and 3%
 What is Forecasting?
 Forecasting is the art and science of predicting
future events
 Forecasting may involve taking historical data (such
as past sales) and projecting them into the future
with mathematical model.
 It may be a subjective or an intuitive prediction
 It may be based on demand-driven data, such as
customer plans to purchase, and projecting them
into the future
 the forecast may involve a combination of these,
that is, a mathematical model adjusted by a
manager’s good judgment.
 What is Forecasting?
 Process of predicting a future event
 Underlying basis of all business decisions
 Production
 Inventory
 Personnel
 Facilities and more
 Forecasting is used to make informed decisions
 Match supply to demand.
 Two important aspects of demand forecasting:
 Level of demand
 Degree of accuracy
 The Realities of Forecast
 Forecasts are seldom perfect
 Most techniques assume an underlying stability
in the system
 Product family and aggregated forecasts are
more accurate than individual product forecasts
 Forecast accuracy decreases as time horizon
increases
 Forecasting Time Horizons
 Short-range forecast
 Up to 1 year, generally less than 3 months
 Purchasing, job scheduling, workforce levels,
job assignments, production levels
 Medium-range forecast
 3 months to 3 years
 Sales and production planning, budgeting
 Long-range forecast
 3+ years
 New product planning, facility location,
research and development
 Influence of Product Life Cycle
Introduction – Growth – Maturity – Decline
 Introduction and growth require longer forecasts
than maturity and decline
 As product passes through life cycle, forecasts are
useful in projecting
 Staffing levels
 Inventory levels
 Factory capacity
Elements of a Good Forecast

Timely

Reliable Accurate

Written

Cost Effective
 Steps in the Forecasting Process
 Determine the use of the forecast
 Select the items to be forecasted
 Determine the time horizon of the forecast
 Select the forecasting model
 Gather the data needed to make the
forecast
 Make the forecast
 Validate and implement the results
 Forecasting Approaches
Qualitative Methods
• Used when little data exist e.g. New products,
• Involves intuition, experience, e.g., forecasting sales on
Internet
① Jury of executive opinion (Pool opinions of high-level experts,
sometimes augment by statistical models)
② Delphi method (Panel of experts, queried iteratively until
consensus is reached)
③ Sales force composite (Estimates from individual salespersons
are reviewed for reasonableness, then aggregated)
④ Consumer Market Survey (Ask the customer)
 Quantitative Approaches
 Used when situation is ‘stable’ and historical data exist
 Existing products
 Involves mathematical techniques
 e.g., forecasting sales of LCD televisions
1. Naive approach
2. Moving averages
time-series
3. Exponential models
smoothing
4. Trend projection
5. Linear regression associative
model
 Time Series Forecasting
 Set of evenly spaced numerical data
 Obtained by observing response variable at regular time
periods
 Forecast based only on past values, no
other variables important
 Assumes that factors influencing past and present will
continue influence in future
 A time series typically has 4 components:
 Trend,
 Seasonality,
 Cycles,
 random variation.
Trend Component
 Persistent, overall upward or downward pattern
 Changes due to population, technology, age, culture, etc.
 Typically several years duration
Seasonal Component
 Regular pattern of up and down fluctuations
 Due to weather, customs, etc.
 Occurs within a single year
Cyclical Component
 Repeating up and down movements
 Affected by business cycle, political, and economic factors
 Multiple years duration
Random Component
 Erratic, unsystematic, ‘residual’ fluctuations
 Due to random variation or unforeseen events
 Short duration and nonrepeating
 Components of Demand
Trend
component
Demand for product or service

Seasonal peaks

Actual demand
line

Average demand
over 4 years

Random variation
| | | |
1 2 3 4
Time (years)
Figure 4.1
 Naive Approach
 Assumes demand in next period is the same
as demand in most recent period
 e.g., If January sales were 68, then February sales
will be 68
 Simple to use
 Virtually no cost
 Quick and easy to prepare
 Data analysis is nonexistent
 Sometimes cost effective and efficient
 Can be good starting point
 Moving Average Method
 MA is a series of arithmetic means
 A MA forecast uses a number of the most recent
actual data values in generating a forecast.
 Used if little or no trend
 Used often for smoothing
 Provides overall impression of data over time

∑ demand in previous n periods

Moving average = n
 Moving Average Example

Calculate a 3-month moving-average forecast,

Actual
Month Shed Sales

January 10
February 12
March 13
April 16
May 19
June 23
July 26
 Moving Average Example
Actual 3-Month
Month Shed Sales Moving Average

January 10
February 12
March 13
April 16 (10 + 12 + 13)/3 = 11.66
May 19 (12 + 13 + 16)/3 = 13.66
June 23 (13 + 16 + 19)/3 = 16
July 26 (16 + 19 + 23)/3 = 19.33
 Simple Moving Average
Actual
MA5
47
45
43
41
39
37 MA3
35
1 2 3 4 5 6 7 8 9 10 11 12
 Weighted Moving Average
 Used when some trend might be present
 Older data usually less important
 Weights based on experience and intuition
∑ (weight for period n)
x (demand in period n)
Weighted Moving
Average = ∑ weights

Ft = WMAn= wnAt-n + … wn-1At-2 + w1At-1

 Weighted Moving Average Example
Period Weights Applied
Last month 3
Two months ago 2
Three months ago 1
Sum of weights 6

Actual
Month Shed Sales

January 10
February 12
March 13
April 16
May 19
June 23
July 26
 Weighted Moving Average Example
Period Weights Applied
Last month 3
Two months ago 2
Three months ago 1
Sum of weights 6

Actual 3-Month Weighted

Month Shed Sales Moving Average

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
July 26 [(3 x 23) + (2 x 19) + (16)]/6 = 20.5
Weighted Moving Average Example
Period Demand Weight

1 42

2 40 10%

3 43 20%

4 40 30%

5 41 40%

Calculate a four year weighted moving Average for year 6.

Weighted Moving Average Example
Period Demand Weight

1 42

2 40 10%

3 43 20%

4 40 30%

5 41 40%

Calculate a four year weighted moving Average for year 6.

 Potential Problems With
Moving Average
Both simple and weighted moving averages are
effective in smoothing out sudden fluctuations in
the demand pattern to provide stable estimates.
 Moving averages do, however, present three
problems
 Increasing n smooth the forecast but makes it
less sensitive to changes
 Moving averages cannot pick up trends very well.
 Moving averages require extensive records of past
data.
 Moving Average And
Weighted Moving Average
Weighted
moving
30 – average
25 –
Sales demand

20 – Actual
sales
15 –
Moving
10 – average

5 –
| | | | | | | | | | | |
Figure 4.2
J F M A M J J A S O N D
 Exponential Smoothing
• Form of weighted moving average
– Weights decline exponentially
– Most recent data weighted most
– Uses most recent period’s actual and forecast data

• Requires smoothing constant ()

– Ranges from 0 to 1
– Subjectively chosen

• A-F is the error term,  is the % feedback

• Involves little record keeping of past data
 Exponential Smoothing
st = Last period’s forecast
+ a (Last period’s actual demand
– Last period’s forecast)

Ft = Ft – 1 + a(At – 1 - Ft – 1)
where Ft = new forecast
Ft – 1 = previous forecast
a = smoothing (or weighting)
constant (0 ≤ a ≤ 1)
 Exponential Smoothing Example

In January, a car dealer predicted February

demand for 142 Ford Mustangs. Actual February
demand was 153 autos. Using a smoothing
constant chosen by management of α = .20, the
dealer wants to forecast March demand using the
exponential smoothing model.
 Exponential Smoothing Example
In January, a car dealer predicted February demand for 142 Ford
Mustangs. Actual February demand was 153 autos. Using a
smoothing constant chosen by management of α = .20, the dealer
wants to forecast March demand using the exponential smoothing
model.
Predicted demand(t-1) = 142 Ford Mustangs
Actual demand (t-1)= 153
Smoothing constant a = .20

New forecast = 142 + .2(153 – 142)

= 142 + 2.2
= 144.2 ≈ 144 cars
 Impact of Different 

225 –
Actual a = .5
demand
200 –
Demand

175 –
a = .1
| | | | | | | | |
150 –
1 2 3 4 5 6 7 8 9
Quarter
 Impact of Different 

225 –
Actual a = .5
 Chose high of 
values
demand
when
200 – underlying average
Demand

is likely to change
175
Choose
– low values of 
when underlying average a = .1
is stable
| | | | | | | | |
150 –
1 2 3 4 5 6 7 8 9
Quarter
 Measuring Forecast Error
• The objective is to obtain the most accurate forecast no
matter the technique
• We generally do this by selecting the model that gives
us the lowest forecast error

Forecast error = Actual demand - Forecast value

= At - Ft
Three commonly used measures for summarizing
Forecasting errors are
• Mean absolute deviation (MAD)
• Mean squared error (MSE)
• Mean absolute percent error (MAPE)
 MAD, MSE, and MAPE

 Actual  forecast
MAD =
n
2
 ( Actual  forecast)
MSE =
n

( Actual  forecast / Actual)*100)

MAPE =
(
n
 Example
• During the past 8 quarters, the Port of Baltimore has
unloaded large quantities of grain from ships. The
Port’s Operations Manager wants to test the
forecasting method exponential smoothing to see
how well the this method works in predicting
tonnage unloaded.
• He guesses that the forecast of grain unloaded in the
first quarter was 175 tons.
 Comparison of Forecast Error
Rounded Absolute Rounded Absolute
Actual Forecast Deviation Forecast Deviation
Tonnage with for with for
Quarter Unloaded a = .10 a = .10 a = .50 a = .50
1 180 175 5.00 175 5.00
2 168 175.5 7.50 177.50 9.50
3 159 174.75 15.75 172.75 13.75
4 175 173.18 1.82 165.88 9.12
5 190 173.36 16.64 170.44 19.56
6 205 175.02 29.98 180.22 24.78
7 180 178.02 1.98 192.61 12.61
8 182 178.22 3.78 186.30 4.30
82.45 98.62
 Comparison of Forecast Error
∑ |deviations|
Rounded Absolute Rounded Absolute
MADActual
= Forecast Deviation Forecast Deviation
Tonnagen with for with for
Quarter Unloaded a = .10 a = .10 a = .50 a = .50
1
For a =
180
.10 175 5.00 175 5.00
2 168 = 82.45/8
175.5 = 10.31
7.50 177.50 9.50
3 159 174.75 15.75 172.75 13.75
4 For a 175
= .50 173.18 1.82 165.88 9.12
5 190 173.36 16.64 170.44 19.56
6 205 = 98.62/8
175.02 = 12.33
29.98 180.22 24.78
7 180 178.02 1.98 192.61 12.61
8 182 178.22 3.78 186.30 4.30
82.45 98.62
 Comparison of Forecast Error
∑ (forecast errors)
Rounded
2
Absolute Rounded Absolute
MSE = Actual Forecast Deviation Forecast Deviation
n
Tonnage with for with for
Quarter Unloaded a = .10 a = .10 a = .50 a = .50
1
For a 180
= .10 175 5.00 175 5.00
2 = 1,526.54/8
168 175.5 = 190.82
7.50 177.50 9.50
3 159 174.75 15.75 172.75 13.75
4 For a 175
= .50 173.18 1.82 165.88 9.12
5 190 173.36 16.64 170.44 19.56
6 = 1,561.91/8
205 175.02 = 195.24
29.98 180.22 24.78
7 180 178.02 1.98 192.61 12.61
8 182 178.22 3.78 186.30 4.30
82.45 98.62
MAD 10.31 12.33
 Comparison of Forecast Error
n
∑100|deviation
Rounded i|/actualiRounded
Absolute Absolute
MAPE =Actual i=1 Forecast Deviation Forecast Deviation
Tonnage with for with for
Quarter n
Unloaded a = .10 a = .10 a = .50 a = .50
1
For a
180
= .10 175 5.00 175 5.00
2 168 = 175.5
44.75/8 = 5.59%
7.50 177.50 9.50
3 159 174.75 15.75 172.75 13.75
4 For a
175= .50 173.18 1.82 165.88 9.12
5 190 173.36 16.64 170.44 19.56
6 205 = 54.05/8
175.02 =29.98
6.76% 180.22 24.78
7 180 178.02 1.98 192.61 12.61
8 182 178.22 3.78 186.30 4.30
82.45 98.62
MAD 10.31 12.33
MSE 190.82 195.24
 Comparison of Forecast Error
Rounded Absolute Rounded Absolute
Actual Forecast Deviation Forecast Deviation
Tonnage with for with for
Quarter Unloaded a = .10 a = .10 a = .50 a = .50
1 180 175 5.00 175 5.00
2 168 175.5 7.50 177.50 9.50
3 159 174.75 15.75 172.75 13.75
4 175 173.18 1.82 165.88 9.12
5 190 173.36 16.64 170.44 19.56
6 205 175.02 29.98 180.22 24.78
7 180 178.02 1.98 192.61 12.61
8 182 178.22 3.78 186.30 4.30
82.45 98.62
MAD 10.31 12.33
MSE 190.82 195.24
MAPE 5.59% 6.76%
Exponential Smoothing with Trend Adjustment
When a trend is present, exponential smoothing must be
modified to respond to trend
• Assume that demand for our product or service has been
increasing by 100 units per month and that we have been
forecasting with α = 0.4 in our exponential smoothing model.
The following table shows a severe lag in the 2nd, 3rd, 4th, and
5th months, even when our initial estimate for month 1 is
perfect:
 Exponential Smoothing with Trend Adjustment

Forecast Exponentially Exponentially

including (FITt) = smoothed (Ft) + smoothed (Tt)
trend Forecast trend

Ft = a(At - 1) + (1 - a)(Ft - 1 + Tt - 1)

Tt = b(Ft - Ft - 1) + (1 - b)Tt - 1
So the three steps to compute a trend-
adjusted forecast are:
Step 1: Compute Ft, the exponentially
smoothed forecast for period t
Step 2: Compute the smoothed trend, Tt
Step 3: Calculate the forecast including
trend, FITt , by the formula FITt = Ft + Tt
 EXAMPLE
• A Portland manufacturer wants to forecast the
demand for a pollution-control equipment.
Past data shows that there is an increasing
trend. The company assumes the initial
forecast for month 1 was 11 units and the
trend over that period was 2 units.
• α = 0.2 β =0.4
 Exponential Smoothing with
Trend Adjustment Example
Forecast
Actual Smoothed Smoothed Including
Month(t) Demand (At) Forecast, Ft Trend, Tt Trend, FITt
1 12 11 2 13.00
2 17
3 20
4 19
5 24
6 21
7 31
8 28
9 36
10
 Exponential Smoothing with
Trend Adjustment Example
Forecast
Actual Smoothed Smoothed Including
Month(t) Demand (At) Forecast, Ft Trend, Tt Trend, FITt
1 12 11 2 13.00
2 17
3 20
4 19
Step 1: Forecast for Month 2
5 24
6 21
F2 = aA1 + (1 - a)(F1 + T1)
7 31
8 28 F2 = (.2)(12) + (1 - .2)(11 + 2)
9 36 = 2.4 + 10.4 = 12.8 units
10
 Exponential Smoothing with
Trend Adjustment Example
Forecast
Actual Smoothed Smoothed Including
Month(t) Demand (At) Forecast, Ft Trend, Tt Trend, FITt
1 12 11 2 13.00
2 17 12.80
3 20
4 19
Step 2: Trend for Month 2
5 24
6 21
T2 = b(F2 - F1) + (1 - b)T1
7 31
8 28 T2 = (.4)(12.8 - 11) + (1 - .4)(2)
9 36 = .72 + 1.2 = 1.92 units
10
 Exponential Smoothing with
Trend Adjustment Example
Forecast
Actual Smoothed Smoothed Including
Month(t) Demand (At) Forecast, Ft Trend, Tt Trend, FITt
1 12 11 2 13.00
2 17 12.80 1.92
3 20
4 19
Step 3: Calculate FIT for Month 2
5 24
6 21
FIT2 = F2 + T2
7 31
8 28 FIT2 = 12.8 + 1.92
9 36 = 14.72 units
10
 Exponential Smoothing with
Trend Adjustment Example
Forecast
Actual Smoothed Smoothed Including
Month(t) Demand (At) Forecast, Ft Trend, Tt Trend, FITt
1 12 11 2 13.00
2 17 12.80 1.92 14.72
3 20 15.18 2.10 17.28
4 19 17.82 2.32 20.14
5 24 19.91 2.23 22.14
6 21 22.51 2.38 24.89
7 31 24.11 2.07 26.18
8 28 27.14 2.45 29.59
9 36 29.28 2.32 31.60
10 32.48 2.68 35.16
 Exponential Smoothing with
Trend Adjustment Example
35 –
Actual demand (At)
30 –
Product demand

25 –

20 –

15 –

10 –
Forecast including trend (FITt)
with  = .2 and  = .4
5 –

0 – | | | | | | | | |
1 2 3 4 5 6 7 8 9
Figure 4.3
Time (month)
 Trend Projections
• The trend projection technique fits a
trend line to a series of historical data
points, and then projects the line for
medium-to long-range forecasts.
• Though there are several mathematical
trend equations (e.g. quadratic and
exponential) available, we will restrict
our discussion to a linear trend (simple
linear regression only).
 

55
 Trend Projections
• Linear trends can be found using the least squares
technique
• This approach results in a straight line that minimizes
the sum of the squares of the vertical differences or
deviations from the line to each of the actual
observations.

^
y = a + bx
^ where y = computed value of the variable to be
predicted (dependent variable)
a = y-axis intercept
b = slope of the regression line
x = the independent variable (which in this case is
time)
Values of Dependent Variable Least Squares Method

Actual observation Deviation7

(y-value)

Deviation5 Deviation6

Deviation3

Deviation4

Deviation1
(error) Deviation2
Trend line, y^ = a + bx

Time period Figure 4.4

Values of Dependent Variable Least Squares Method

Actual observation Deviation7

(y-value)

Deviation5 Deviation6

Deviation3
Least squares method minimizes the sum
of the squared errors (deviations)
Deviation4

Deviation1
(error) Deviation2
Trend line, y^ = a + bx

Time period Figure 4.4

 Least Squares Method
Equations to calculate the regression variables
^
y = a + bx
Sxy - nxy
b=
Sx2 - nx2

a = y - bx
 Least Squares Example
Time Electrical Power
Year Period (x) Demand (megawatt) x2 xy
2006 1 74 1 74
2007 2 79 4 158
2008 3 80 9 240
2009 4 90 16 360
2010 5 105 25 525
2011 6 142 36 852
2012 7 122 49 854
∑x = 28 ∑y = 692 ∑x2 = 140 ∑xy = 3,063
x=4 y = 98.86

∑xy - nxy 3,063 - (7)(4)(98.86)

b= = = 10.54
∑x2 - nx2 140 - (7)(42)

a = y - bx = 98.86 - 10.54(4) = 56.70

 Least Squares Example
Time Electrical Power
Year Period (x) Demand x2 xy
2003 1 74 1 74
2004 2 79 4 158
2005 The trend3 line is 80 9 240
2006 4 90 16 360
^
2007 y5= 56.70 + 10.54x
105 25 525
2008 6 142 36 852
2009 7 122 49 854
∑x = 28 ∑y = 692 ∑x2 = 140 ∑xy = 3,063
x=4 y = 98.86

∑xy - nxy 3,063 - (7)(4)(98.86)

b= = = 10.54
∑x - nx
2 2 140 - (7)(42)

a = y - bx = 98.86 - 10.54(4) = 56.70

 Least Squares Example
Trend line,
160 – ^ y = 56.70 + 10.54x
150 –
140 –
Power demand

130 –
120 –
110 –
100 –
90 –
80 –
70 –
60 –
50 –
| | | | | | | | |
2006 2007 2008 2009 2010 2011 2012 2013 2014
Year
Least Squares Requirements
• We always plot the data to insure a
linear relationship
• We do not predict time periods far
beyond the database
• Deviations around the least squares
line are assumed to be random and
normally distributed.
 Seasonal Variations In Data
• Seasonal variations in data are regular up-and-
down movements in a time series that relate to
recurring events such as weather or holidays.
• Seasonality may be applied to hourly, daily,
weekly, monthly, or other recurring patterns.
• Understanding seasonal variations is important
for capacity planning in organizations that
handle peak loads.
• The presence of seasonality makes adjustments in
trend-line forecasts necessary.
• Seasonality is expressed in terms of the amount
that actual values differ from average values in the
time series.
• Analyzing data in monthly or quarterly terms
usually makes it easy for a statistician to spot
seasonal patterns.
• The multiplicative seasonal model can adjust trend
data for seasonal variations in demand
• Assumption in this section is that trend has been
removed from the data.
 Seasonal Index Example
Demand Average Average Seasonal
Month 2010 2011 2012 2010-2012 Monthly Index
Jan 80 85 105 90 94
Feb 70 85 85 80 94
Mar 80 93 82 85 94
Apr 90 95 115 100 94
May 113 125 131 123 94
Jun 110 115 120 115 94
Jul 100 102 113 105 94
Aug 88 102 110 100 94
Sept 85 90 95 90 94
Oct 77 78 85 80 94
Nov 75 72 83 80 94
Dec 82 78 80 80 94
 Seasonal Index Example
Demand Average Average Seasonal
Month 2010 2011 2012 2010-2012 Monthly Index
Jan 80 85 105 90 94 0.957
Feb 70 85 85 80 94
Mar 80 93 Average
82 85 monthly demand
2010-2012 94
AprSeasonal90index 95
= 115
Average 100demand
monthly 94
May 113 125 131 123 94
Jun 110 115= 90/94
120 = .957 115 94
Jul 100 102 113 105 94
Aug 88 102 110 100 94
Sept 85 90 95 90 94
Oct 77 78 85 80 94
Nov 75 72 83 80 94
Dec 82 78 80 80 94
 Seasonal Index Example
Demand Average Average Seasonal
Month 2010 2011 2012 2010-2012 Monthly Index
Jan 80 85 105 90 94 0.957
Feb 70 85 85 80 94 0.851
Mar 80 93 82 85 94 0.904
Apr 90 95 115 100 94 1.064
May 113 125 131 123 94 1.309
Jun 110 115 120 115 94 1.223
Jul 100 102 113 105 94 1.117
Aug 88 102 110 100 94 1.064
Sept 85 90 95 90 94 0.957
Oct 77 78 85 80 94 0.851
Nov 75 72 83 80 94 0.851
Dec 82 78 80 80 94 0.851
 Seasonal Index Example
Demand Average Average Seasonal
Month 2010 2011 2012 2010-2012
Monthly Index
Jan 80 85 105 90 94 0.957
Feb 70 85 85Forecast for 2013
80 94 0.851
Mar 80 93 82 85 94 0.904
Apr 90 95Expected
115 annual demand
100 = 1,200 94 1.064
May 113 125 131 123 94 1.309
Jun 110 115 120 1,200 115 94 1.223
Jul 100 102 Jan 113 12 105x .957 = 96 94 1.117
Aug 88 102 110 100 94 1.064
1,200
Sept 85 90 Feb 95 90
x .851 = 85 94 0.957
12
Oct 77 78 85 80 94 0.851
Nov 75 72 83 80 94 0.851
Dec 82 78 80 80 94 0.851
 Seasonal Index Example
2013 Forecast
140 – 2012 Demand
130 – 2011 Demand
2010 Demand
120 –
Demand

110 –
100 –
90 –
80 –
70 –
| | | | | | | | | | | |
J F M A M J J A S O N D
Time
 Associative Forecasting
Used when changes in one or more independent
variables can be used to predict the changes in the
dependent variable

Most common technique is linear

regression analysis

We apply this technique just as we did in the

time series example
 Associative Forecasting
Forecasting an outcome based on predictor
variables using the least squares technique

^
y = a + bx
^ where y = computed value of the variable to be
predicted (dependent variable)
a = y-axis intercept
b = slope of the regression line
x = the independent variable though to predict the
value of the dependent variable
 Associative Forecasting Example

Nodel Construction Company renovates old homes in West

Bloomfield, Michigan. Over time, the company has found that its
dollar volume of renovation work is dependent on the West
Bloomfield area payroll. Management wants to establish a
mathematical relationship to help predict sales.
Sales Area Payroll
($ millions), y ($ billions), x
2.0 1
3.0 3 4.0 –
2.5 4
2.0 2 3.0 –
Sales
2.0 1
3.5 7 2.0 –

1.0 –
| | | | | | |
0 1 2 3 4 5 6 7
Area payroll
Associative Forecasting Example
Sales, y Payroll, x x2 xy
2.0 1 1 2.0
3.0 3 9 9.0
2.5 4 16 10.0
2.0 2 4 4.0
2.0 1 1 2.0
3.5 7 49 24.5
∑y = 15.0 ∑x = 18 ∑x2 = 80 ∑xy = 51.5

∑xy 51.5 - (6)(3)(2.5)

x = ∑x/6 = 18/6 = 3 b =- nxy = = .25
∑x2 - nx2 80 - (6)(3 )
2

y = ∑y/6 = 15/6 = 2.5 a = y - bx = 2.5 - (.25)(3) = 1.75

 Associative Forecasting Example
^
y = 1.75 + .25x Sales = 1.75 + .25(payroll)

If payroll next year is

4.0 –
estimated to be $6 3.25
billion, then: 3.0 –
Nodel’s sales

Sales = 1.75 + .25(6) 2.0 –

Sales = $3,250,000 1.0 –

| | | | | | |
0 1 2 3 4 5 6 7
Area payroll
Standard Error of the Estimate
 A forecast is just a point estimate of a
future value
 This point is 4.0 –
actually the 3.25
3.0 –
mean of a Nodel’s sales

probability 2.0 –

distribution 1.0 –
| | | | | | |
0 1 2 3 4 5 6 7
Area payroll
Figure 4.9
 Standard Error of the Estimate
• To measure the accuracy of the regression estimates, we
must compute the standard error of the estimate , Sy,x
• This computation is called the standard deviation of the
regression:
• It measures the error from the dependent variable, y , to
the regression line, rather than to the mean.

∑y2 - a∑y - b∑xy

Sy,x =
n-2
We use the standard error to set up prediction
intervals around the point estimate
 Standard Error of the Estimate
Nodel’s VP of operations now wants to know the error
associated with the regression line computed in the
previous Example
∑y2 - a∑y - b∑xy 39.5 - 1.75(15) - .25(51.5)
Sy,x = =

n-2 6-2

Sy,x = .306
4.0 –
3.25
Nodel’s sales 3.0 –
The standard error
of the estimate is 2.0 –
$306,000 in sales
1.0 –
| | | | | | |
0 1 2 3 4 5 6 7
Area payroll
 Prediction Intervals
The ranges or limits of a forecast are estimated by:
Upper limit = Y + t*Sy,x
Lower limit = Y - t*Sy,x
where:
Y = forecasted value (point Estimate)
t = number of standard deviations from the mean of
the distribution to provide a given probability of exceeding
the limits through chance
syx = standard error of the forecast
 Prediction Intervals
Estimate the confidence interval for annual sales of
next year (remember payroll next year is estimated to
be $6 billion) so that the probability that the actual
sales exceeding these limits will be approximately equal
to 0.05.
Sales = 1.75 + .25(6)
Sales = $3,250,000
 Prediction Intervals

Step 1: Compute the standard error of the

forecasts, syx.
Step 2: Determine the appropriate value for t.
n = 6, so degrees of freedom = n – 2 = 4
level of significance = =.05
The student’s t Distribution table shows:
(tα/2=0.025) ) =2.78
 Prediction Intervals
Step 3: Compute upper and lower limits.
Upper limit = $3,250,000 + 2.78 * $306,000
=$4,100,680
Lower limit = $3,250,000 - 2.78 * $306,000 =
$2,399,320
We are 95% confident the actual sales for next year will
be between $2,399,320 and $4,100,680.
 Correlation
• How strong is the linear relationship
between the variables?
• Correlation does not necessarily imply
causality!
• Coefficient of correlation, r, measures
degree of association
– Values range from -1 to +1
Correlation Coefficient
nSxy - SxSy
r=
[nSx2 - (Sx)2][nSy2 - (Sy)2]
 Correlation Coefficient
y y

nSxy - SxSy
r=
(a) Perfect positive [nSx
x
2
- (Sx)2][nSy(b)
2
- Positive
(Sy)2] x
correlation: correlation:
r = +1 0<r<1

y y

(c) No correlation: x (d) Perfect negative x

r=0 correlation:
r = -1
 Correlation
 Coefficient of Determination, r2,
measures the percent of change in y
predicted by the change in x
 Values range from 0 to 1
 Easy to interpret

For the Nodel Construction example:

r = .901
r2 = .81
 Multiple Regression Analysis
If more than one independent variable is to be used
in the model, linear regression can be extended to
multiple regression to accommodate several
independent variables
^ y=a+b x +b x …
1 1 2 2

Computationally, this is quite complex and

generally done on the computer
 Multiple Regression Analysis
In the Nodel example, including interest rates in the
model gives the new equation:
^
y = 1.80 + .30x1 - 5.0x2

An improved correlation coefficient of r = .96 means this

model does a better job of predicting the change in
construction sales

Sales = 1.80 + .30(6) - 5.0(.12) = 3.00

Sales = $3,000,000
 Monitoring and Controlling
Forecasts
Tracking Signal
• Measures how well the forecast is predicting
actual values
• Ratio of cumulative forecast errors to mean
absolute deviation (MAD)
– Good tracking signal has low values
– If forecasts are continually high or low, the
forecast has a bias error
 Monitoring and Controlling
Forecasts

Tracking Cumulative error

signal =
MAD

∑(Actual demand in
period i -
Forecast demand
Tracking in period i)
=
signal (∑|Actual - Forecast|/n)
 Tracking Signal
Signal exceeding limit
Tracking signal
Upper control limit
+

Acceptable
0 MADs range

– Lower control limit

Time
 Tracking Signal Example
Carlson’s Bakery wants to evaluate performance of its
croissant forecast. Develop a tracking signal for the
forecast and see if it stays within acceptable limits,
which we define as ±4 MADs
Cumulative
Absolute Absolute
Actual Forecast Cumm Forecast Forecast
Qtr Demand Demand Error Error Error Error MAD
1 90 100 -10 -10 10 10 10.0
2 95 100 -5 -15 5 15 7.5
3 115 100 +15 0 15 30 10.0
4 100 110 -10 -10 10 40 10.0
5 125 110 +15 +5 15 55 11.0
6 140 110 +30 +35 30 85 14.2
 Tracking Signal Example
Cumulative
Tracking Absolute Absolute
Signal Forecast
Actual Cumm Forecast Forecast
Qtr (Cumm Demand
Demand Error/MAD)
Error Error Error Error MAD
1 90-10/10 =100
-1 -10 -10 10 10 10.0
2 95-15/7.5 100
= -2 -5 -15 5 15 7.5
3 1150/10 = 0100 +15 0 15 30 10.0
4 100-10/10 =110
-1 -10 -10 10 40 10.0
5 125+5/11 =110
+0.5 +15 +5 15 55 11.0
6 140+35/14.2110= +2.5 +30 +35 30 85 14.2

The variation of the tracking signal between -2.0 and +2.5 is

within acceptable limits
 Adaptive Smoothing
• Adaptive smoothing refers to computer
monitoring of tracking signals and self
adjustment if a signal exceeds its limit.
• In exponential smoothing, the coefficients are
first selected based on values that minimize
error forecasts and then adjusted accordingly
whenever the computer notes an errant
tracking signal. This is called adaptive
smoothing.
 Use of Computer in Forecasting

• Numerous software packages such as SAS,

SPSS, BIOMED, SYSTEB, Minitab etc. are
readily available to handle time series and
causal forecasting projections.
• Even spreadsheet software such as Lotus1-2-3,
can effectively manage small-to-medium
range forecasting problems.
Forecasting in the Service Sector

 Presents unusual challenges

 Special need for short term records
 Needs differ greatly as function of
industry and product
 Holidays and other calendar events
 Unusual events