Chapter 2:

Forecasting

COMPILED BY: Tewodros B.

Harambe University
OM is mostly proactive not reactive
It involves structured planning activities

HARAMBE UNIVERSITY COLLE

GE
 Objectives
 Give the fundamental rules of forecasting
 Calculate a forecast using a moving average,
weighted moving average, and exponential
smoothing
 Calculate the accuracy of a forecast

HARAMBE UNIVERSITY COLLE

GE
 What is forecasting?
 Forecasting is a tool used for predicting
 future demand based on
 past demand information.

 Process of predicting a
future event
 Underlying basis ??
of all business decisions
 Production
 Inventory
 Personnel
 Facilities

HARAMBE UNIVERSITY COLLE

GE
 Why is forecasting important?
Demand for products and services is usually uncertain.
Forecasting can be used for…
• Strategic planning (long range planning)
• Finance and accounting (budgets and cost controls)
• Marketing (future sales, new products)
• Production and operations
 Forecasts affect decisions and activities throughout an organization
 Accounting, finance

 Human resources

 Marketing

 MIS

 Operations

 Product / service design

HARAMBE UNIVERSITY COLLE
GE
 Why is forecasting important?
Accounting Cost/profit estimates

Finance Cash flow and funding

Human Resources Hiring/recruiting/training

Marketing Pricing, promotion, strategy

MIS IT/IS systems, services

Operations Schedules, MRP, workloads

Product/service design New products and services

What is forecasting all about?

Demand for Mercedes E Class We try to predict the

future by looking back
at the past

Predicted
demand
looking
Time back six
Jan Feb Mar Apr May Jun Jul Aug months
Actual demand (past sales)
Predicted demand

HARAMBE UNIVERSITY COLLE

GE
 Elements of a Good Forecast

Timely

Reliable Accurate

se
ul u
f to
ing Written sy
an Ea
e
M
HARAMBE UNIVERSITY COLLE
GE
 Some general characteristics of forecasts
Forecasts are always wrong/Forecasts are
seldom perfect
Forecasts are more accurate for groups or
families of items/Product family and
aggregated forecasts are more accurate
than individual product forecasts
Forecasts are more accurate for shorter
time periods/Forecast accuracy decreases as
time horizon for forecasts increases
Ex. I can forecast this year’s class
average better than next year’s class
average HARAMBE UNIVERSITY COLLE
 GE
Steps in the Forecasting Process

“The forecast”

Step 6 Monitor the forecast

Step 5 Prepare the forecast
Step 4 Gather and analyze data
Step 3 Select a forecasting technique
Step 2 Establish a time horizon
Step 1 Determine purpose of forecast

HARAMBE UNIVERSITY COLLE

GE
 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
HARAMBE UNIVERSITY COLLE
GE
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

HARAMBE UNIVERSITY COLLE

GE
 Product Life Cycle
Introduction Growth Maturity Decline
Product design Forecasting critical Standardization Little product
and development Product and Fewer product differentiation
critical process reliability changes, more Cost
Frequent product Competitive minor changes minimization
and process design product Optimum capacity Overcapacity
changes improvements and in the industry
Increasing stability
Short production options of process Prune line to
runs Increase capacity eliminate items
Long production
OM Strategy/Issues

High production Shift toward runs not returning

costs product focus good margin
Product
Limited models Enhance improvement and Reduce
Attention to distribution cost cutting capacity
quality

HARAMBE UNIVERSITY
COLLEGE
Key issues in forecasting

1. A forecast is only as good as the information included in the

forecast (past data)
2. History is not a perfect predictor of the future (i.e.: there is
no such thing as a perfect forecast)

REMEMBER: Forecasting is based on the assumption

that the past predicts the future! When forecasting, think
carefully whether or not the past is strongly related to
what you expect to see in the future…
HARAMBE UNIVERSITY COLLE
GE
Example: Mercedes E-class vs. M-class Sales
Month E-class Sales M-class Sales
Jan 23,345 -
Feb 22,034 -
Mar 21,453 -
Apr 24,897 -
May 23,561 -
Jun 22,684 -
Jul ? ?

Question: Can we predict the new model M-class sales based on

the data in the the table?

Answer: Maybe... We need to consider how much the two

markets have in common
HARAMBE UNIVERSITY COLLE
GE
What should we consider when looking at
past demand data?

• Trends

• Seasonality

• Cyclical elements

• Autocorrelation

• Random variation

HARAMBE UNIVERSITY COLLE

GE
How should we pick our forecasting model?

1. Data availability
2. Time horizon for the forecast
3. Required accuracy
4. Required Resources

HARAMBE UNIVERSITY COLLE

GE
 Forecasting Approaches
Qualitative Methods
 Used when little data exist
 New products
 New technology
 Involves intuition, experience
 e.g., forecasting sales on Internet

HARAMBE UNIVERSITY COLLE

GE
 Qualitative Methods

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)

HARAMBE UNIVERSITY COLLE

GE
 Quantitative Approaches
 Used when situation is ‘stable’ and
historical data exist
 Existing products
 Current technology
 Involves mathematical techniques
 e.g., forecasting sales of LCD televisions

HARAMBE UNIVERSITY COLLE

GE
 Quantitative Approaches

1. Naive approach
2. Moving averages time-series
models
3. Exponential
smoothing associative
model
4. Trend projection
5. Linear regression

HARAMBE UNIVERSITY COLLE

GE
 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

HARAMBE UNIVERSITY COLLE

GE
 Time Series Components

Trend Cyclical

Seasonal Random

HARAMBE UNIVERSITY COLLE

GE
 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)
HARAMBE UNIVERSITY COLLE
GE
 Trend Component
 Persistent, overall upward or
downward pattern
 Changes due to population,
technology, age, culture, etc.
 Typically several years duration

HARAMBE UNIVERSITY COLLE

GE
 Seasonal Component
 Regular pattern of up and down
fluctuations
 Due to weather, customs, etc.
 Occurs within a single year
Number of
Period Length Seasons
Week Day 7
Month Week 4-4.5
Month Day 28-31
Year Quarter 4
Year Month 12
Year Week 52
HARAMBE UNIVERSITY COLLE
GE
 Cyclical Component
 Repeating up and down movements
 Affected by business cycle, political, and
economic factors
 Multiple years duration

0
HARAMBE UNIVERSITY 5
COLLE 10 15 20
GE
 Random Component
 Erratic, unsystematic, ‘residual’ fluctuations
 Due to random variation or unforeseen events
 Short duration
and non repeating

MCOLLE
HARAMBE UNIVERSITY T W T F
GE
Forecasting Methods for Time Series model
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
 Sometimes cost effective and efficient
 Can be good starting point

HARAMBE UNIVERSITY COLLE

GE
 2. Moving Average
A n-month moving average is the sum of the observed values
during the past n months divided by n.
∑ demand in previous n periods
Moving average =
n

Moving Average Method n =3

Month Sales Forecast Formula
1 100
2 80
3 90
4 110 90.00 =(100+80+90)/3
5 100 93.33 =(80+90+110)/3
6 110 100.00 =(90+110+100)/3
7 95 106.67 =(110+100+110)/3
8 115 101.67 =(100+110+95)/3
9 120 106.67 =(110+95+115)/3
10 90 110.00 =(95+115+120)/3
11 105 108.33 =(115+120+90)/3
12 110 105.00 =(120+90+105)/3
HARAMBE UNIVERSITY COLLE
GE
Time series: simple moving average

In the simple moving average models the forecast value is

At + At-1 + … + At-n
Ft+1 =
n

t is the current period.

Ft+1 is the forecast for next period
n is the forecasting horizon (how far back we look),
A is the actual sales figure from each period.
HARAMBE UNIVERSITY COLLE
GE
 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

HARAMBE UNIVERSITY COLLE

GE
Time series: weighted moving average
We may want to give more importance to some of the data…

Ft+1 = wt At + wt-1 At-1 + … + wt-n At-n

wt + wt-1 + … + wt-n = 1

t is the current period.

Ft+1 is the forecast for next period
n is the forecasting horizon (how far back we look),
A is the actual sales figure from each period.
w is the importance (weight) we give to each period
HARAMBE UNIVERSITY COLLE
GE
How do we choose weights?

1. Depending on the importance that we feel past data has

2. Depending on known seasonality (weights of past data
can also be zero).

WMA is better than SMA

because of the ability to
vary the weights!

HARAMBE UNIVERSITY COLLE

GE
 Weights Applied Period

Weighted Moving Average

3
2
Last month
Two months ago
1 Three months ago
6 Sum of weights

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 = 121/6
May 19 [(3 x 16) + (2 x 13) + (12)]/6 = 141/3
June 23 [(3 x 19) + (2 x 16) + (13)]/6 = 17
July 26 [(3 x 23) + (2 x 19) + (16)]/6 = 201/2
HARAMBE UNIVERSITY COLLE
GE
 Weighted average

Month Demand Compute a weighted average forecast using a

weight of 0.4 for the most recent period, 0.3
1 42 for the next most recent, 0.2 for the next and
2 40 0.1 for the next.
3 43
Continuing with the data on the left
4 40 F(6) = .40(41)+.30(40)+.20(43)+.10(40)=41.0
5 41 If the actual demand for period 6 is 39,
6 39 F(7) = .40(39)+.30(41)+.20(40)+.10(43)=40.2

 The weighted average is more reflective of

the most recent occurrences.

HARAMBE UNIVERSITY COLLE

GE
Why do we need the WMA models?

Because of the ability to give more importance to what

happened recently, without losing the impact of the past.

Demand for Mercedes E-class Actual demand (past sales)

Prediction when using 6-month SMA
Prediction when using 6-months WMA

For a 6-month
SMA, attributing
equal weights to all
past data we miss
Time the downward trend
Jan Feb Mar Apr May Jun Jul Aug
HARAMBE UNIVERSITY COLLE
GE
 3. Exponential Smoothing (ES)

Main idea: The prediction of the future depends mostly on the

most recent observation, and on the error for the latest forecast.

Smoothing Denotes the importance

constant of the past error
alpha α

HARAMBE UNIVERSITY COLLE

GE
Exponential smoothing: the method

Assume that we are currently in period t. We calculated the

forecast for the last period (Ft-1) and we know the actual demand
last period (At-1) …

Ft Ft 1  ( At  1  Ft  1 )

The smoothing constant α expresses how much our forecast will

react to observed differences…
If α is low: there is little reaction to differences.
If α is high: there is a lot of reaction to differences.
HARAMBE UNIVERSITY COLLE
GE
 Exponential Smoothing the method

Forecast error:=Actual – Forecast =A(t-1)-F(t-1)

Ft  Ft  1   ( At  1  Ft  1 )
Forecast today=Forecast yesterday+(alpha)*(Forecast error yesterday)
Each new forecast is equal to the previous forecast plus a percentage of
the previous error.
Today’s forecast
Depends on yesterday’s (time-wise dependence, strong memory)
But it has to be corrected by forecast error
Therefore, we should give more weight to the more recent time
periods when forecasting.
 Alpha = smoothing constant = percentage of the forecast error.
HARAMBE UNIVERSITY COLLE
GE
 Exponential Smoothing
t= 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)

HARAMBE UNIVERSITY COLLE

GE
 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

HARAMBE UNIVERSITY
COLLEGE
 Exponential Smoothing (Example)
Example: F(3) = F(2) + α*{(A(2) – F(2)} = 100 + 0.2*(80 – 100) = 96.
Exponential Smoothing Method (sales are actual sales indicated by A in equation)
alpha = 0.2
Month Sales Forecast Comment and Calculation
Forecast for period 1 should be available before
1 100 100 starting the calculations. If it is not given then set it
equal to the sales of period 1.
2 80 100.00
3 90 96.00
4 110 94.80
5 100 97.84
6 110 98.27
7 95 100.62
8 115 99.50
9 120 102.60
10 90 106.08
11 105 102.86
12 110 103.29
HARAMBE UNIVERSITY COLLE
GE
 Example of Exponential Smoothing
Forecasts made in a period and the period has the same color
Period Actual Forecast withAlpha
Error with
= 0.1
Forecast withError with
1 42 Alpha=0.1 Alpha=0.1 Alpha=0.4 Alpha=0.4
2 40 42 -2.00 42 -2
3 43 41.8 1.20 41.2 1.8
4 40 41.92 -1.92 41.92 -1.92
5 41 41.73 -0.73 41.15 -0.15
6 39 41.66 -2.66 41.09 -2.09
7 46 41.39 4.61 40.25 5.75
8 44 41.85 2.15 42.55 1.45
9 45 42.07 2.93 43.13 1.87
10 38 42.36 -4.36 43.88 -5.88
11 40 41.92 -1.92 41.53 -1.53
12 41.73 40.92

Ft At  1  (1   ) Ft  1
HARAMBE UNIVERSITY COLLE
GE
 Picking a Smoothing Constant:
Responsiveness vs. Smoothing
 The quickness of forecast adjustment to error is determined by the
smoothing constant.
 The closer the alpha is to zero, the slower the forecast will be to
adjust to forecast errors.
 Conversely, the closer the value of alpha is to 1.00, the greater the
responsiveness to the actual observations and the less the
smoothing
 Select a smoothing constant that balances the benefits of
responding to real changes if and when they occur.

Ft Ft  1   ( At  1  Ft  1 ) At  1  (1   ) Ft  1
Why use exponential smoothing?

1. Uses less storage space for data

2. Extremely accurate
3. Easy to understand
4. Little calculation complexity
5. There are simple accuracy tests

HARAMBE UNIVERSITY COLLE

GE
 4. Techniques for trend

 Develop an equation that will describe trend

 The trend component may be linear or it
may not
 Linear trend:
Y
b is similar to the slope.
However, since it is
calculated with the variability
Yt = a + bt of the data in mind, its
formulation is not as
straight-forward as our usual
0 1 2 3 4 5 notion of slope.
tHARAMBE UNIVERSITY COLLE
GE
Common Nonlinear Trends

Parabolic

Exponential

Growth

HARAMBE UNIVERSITY COLLE

GE
Exponential Smoothing with Trend Adjustment
FIT: Forecast including trend
FITt Ft  Tt δ: Trend smoothing constant

Ft FITt  1  α(At  1  FITt  1 )

Tt Tt  1  δ(Ft  FITt  1 )

The idea is that the two effects are decoupled,

(F is the forecast without trend and T is the trend component)
HARAMBE UNIVERSITY COLLE
GE
 Exponential Smoothing with
Trend Adjustment
When a trend is present, exponential
smoothing must be modified to respond to
trend
Forecast Exponentially Exponentially
including (FITt) = smoothed (Ft) + smoothed (Tt)
trend forecast trend

HARAMBE UNIVERSITY COLLE

GE
Example: bottled water at Kroger

At Ft Tt FITt α = 0.8

Jan 1325 1380 -10 1370 δ = 0.5

Feb 1353
Mar 1305
Apr 1275
May 1210
Jun

HARAMBE UNIVERSITY COLLE

GE
Example: bottled water at Kroger

At Ft Tt FITt α = 0.8

Jan 1325 1380 -10 1370 δ = 0.5

Feb 1353 1334 -28 1306

Mar 1305 1344 -9 1334
Apr 1275 1311 -21 1290
May 1210 1278 -27 1251
Jun 1218 -43 1175

HARAMBE UNIVERSITY COLLE

GE
 Exponential Smoothing with
Trend
1400

1350
Actual
1300 a = 0.2

1250 a = 0.8
a = 0.8, d = 0.5
1200

1150
0 1 2 3 4 5 6 7

HARAMBE UNIVERSITY COLLE

GE
 Exercise
 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

HARAMBE UNIVERSITY COLLE

GE
 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

HARAMBE UNIVERSITY COLLE

GE
 Seasonal Variations In Data

The multiplicative
seasonal model can
adjust trend data for
seasonal variations in
demand (jet skis, snow
mobiles)

HARAMBE UNIVERSITY COLLE

GE
 Seasonal Variations In Data
Steps in the process:
1. Find average historical demand for each season
2. Compute the average demand over all seasons
3. Compute a seasonal index for each season
4. Estimate next year’s total demand
5. Divide this estimate of total demand by the number of seasons, then
multiply it by the seasonal index for that season

HARAMBE UNIVERSITY COLLE

GE
 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
HARAMBE UNIVERSITY COLLE
GE
 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
Seasonal90
Apr index 95
= 115 100demand 94
Average monthly
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
HARAMBE UNIVERSITY COLLE
GE
 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
HARAMBE UNIVERSITY COLLE
GE
 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 102Jan 113 12 x .957 = 96
105 94 1.117
Aug 88 102 110 100 94 1.064
1,200
Sept 85 90Feb 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
HARAMBE UNIVERSITY COLLE
GE
 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
HARAMBE UNIVERSITY COLLE
GE
How can we compare across forecasting models?

We need a metric that provides estimation of accuracy

Errors can be:

Forecast Error 1. biased (consistent)

2. random

Forecast error = Difference between actual and forecasted value

(also known as residual)
HARAMBE UNIVERSITY COLLE
GE
 Measuring Accuracy
• We need a way to compare different time series techniques
for a given data set.
• Four common techniques are the: n 
Yi  Y

Mean Absolute Deviation, MAD = 
i 1 n
i

100 n Yi  Ŷi
 Mean Absolute Percent Error, MAPE = 
n i 1 Yi

 The Mean Square Error, MSE = 

n
Yi  Yi 
 2

i 1 n
RMSE  MSE
 Root Mean Square Error.
Where: Y= actual dd, Ŷ= forecasted dd, n=no of observation, i=
a given time period
HARAMBE UNIVERSITY
 Measuring Accuracy, examples
Perio Actual Forecast A-F/ |error| Error [|error|/actual]x 100
d Error 2

1 217 215 2 2 4 .92%

2 213 216
3 216 215
4 210 214
5 213 211
6 219 214
7 216 217
8 212 216
∑ -2 22 76 10.26%
MAD= ∑|e|/n= 22/8 = 2.75

MSE= ∑e2 / n-1 = 76/8 = 9.5

]
MAPE= ∑[|e|/Actual x 100 /n =10.26%/8 = 1.28%
 Measuring Accuracy, examples
Perio Actual Forecast A-F/ |error| Error [|error|/actual]x 100
d Error 2

1 217 215 2 2 4 .92%

2 213 216 -3 3 6 1.14
3 216 215 1 1 1 0.46
4 210 214 -4 4 16 1.90
5 213 211 2 2 4 0.94
6 219 214 5 5 25 2.28
7 216 217 -1 1 1 0.46
8 212 216 -4 4 16 1.89
∑ -2 22 76 10.26%
MAD= ∑|e|/n= 22/8 = 2.75

MSE= ∑e2 / n-1 = 76/8 = 9.5

]
MAPE= ∑[|e|/Actual x 100 /n =10.26%/8 = 1.28%
 Measuring Accuracy…..
• From a computational standpoint, the difference between
these measures is that MAD weights all errors evenly,
MSE weights errors according to their squared values,
and MAPE weights according to relative error.
• Relative to the number of observation. For instance, 10
out of 15 and 10 out of 1000.

HARAMBE UNIVERSITY COLLE

GE
 MFE & MAD:
A Dartboard Analogy

Low MFE & MAD:

The forecast errors

are small &
unbiased

HARAMBE UNIVERSITY COLLE

GE
An Analogy (cont’d)

Low MFE but high

MAD:

On average, the
arrows hit the
bullseye (so much
for averages!)
HARAMBE UNIVERSITY COLLE
GE
MFE & MAD:
An Analogy

High MFE & MAD:

The forecasts
are inaccurate &
biased

HARAMBE UNIVERSITY COLLE

GE
 Key Point

Forecast must be measured for accuracy!

The most common means of doing so is by

measuring the either the mean absolute
deviation or the standard deviation of the
forecast error

HARAMBE UNIVERSITY COLLE

GE
 Which Forecasting Method
Should You Use
 Gather the historical data of what you want to
forecast
 Divide data into initiation set and evaluation set
 Use the first set to develop the models
 Use the second set to evaluate
 Compare the MADs and MFEs of each model

HARAMBE UNIVERSITY COLLE

GE