FORECASTING

Continuation . . .
Quantitative
Methods
• Time Series & Regression

• Time Series  Popular Forecasting Approach in

Operations Management

• Assumption:
• “Patterns” That Occurred in the Past Will Continue to
Occur In the Future

• Patterns
• Random Variation
• Trend
• Seasonality
• Composite
What is a Time Series?

• Obtained by observing response variable at

regular time periods

• Set of evenly spaced numerical data

• Forecast based only on past values

• Assumes that factors influencing past and present will
continue influence in future

• Assumes that factors influencing the past will

continue to influence the future
160 Monthly Champagne
0
Sales
140
0

120
0

100
0

800

600

400
0 1 2 36 6 7 8
200 2 4 48 0 2 4
0
 Trend Component

• Persistent, overall upward or downward

pattern

• Due to population, technology etc.

• Several years duration

Respons
e
Mo., Qtr., © 1984-1994
T/Maker Co.
Yr.
Seasonal Component
• Regular pattern of up & down
fluctuations

• Due to weather, customs etc.

• Occurs within 1 year

Summe
Respons r
e
© 1984-1994
T/Maker Co.

Mo.,
Qtr.
 UK Airline Miles
U.K. Airline Miles

U.K. Airline
Miles
1800
0
Observe:
Increasing trend,
1600
0 Seasonal component.
1400 Random variation.
0
MilesMiles

1200
0
Thousands ofof

1000
0
Thousands

800
0

600
0

400
0

200
0

0
1
4
7
1
0
1
3
1
6
1
9
2
2
2
5
2
8
3
1
3
4
3
7
4
0
4
3
4
6
4
9
5
2
5
5
5
8
6
1
Month
Common Seasonal
Patterns
Period of Pattern “Season” Number of
Length “Seasons”
in Pattern
Week Day 7
Month Week 4–4½
Month Day 28 – 31
Year Quarter 4
Year Month 12
Year Week 52
Cyclical Component
• Repeating up & down movements

• Due to interactions of factors influencing

economy

• Usually 2-10 years duration

Cycle


Response

Mo., Qtr.,
Yr.
Random Component
• Erratic, unsystematic, ‘residual’
fluctuations

• Due to random variation or unforeseen

events
• Union strike
• Hurricane/Cyclone

• Short duration & non-repeating

Forecasting
Steps
Data Collection

Collect
Data Analysis Relevant/Reli
able Data
Be Aware of
Model Selection “Garbage-In,
Garbage Out”

Monitoring
Forecasting
Steps
Data Collection

Data Analysis Plot the Data

Identify
Model Selection Patterns

Monitoring
Forecasting
Steps
Data Collection

Choose Model Appropriate

Data Analysis for Data

Consider Complexity Trade-

Offs
Model Selection Perform
Forecast(s)
Select Model Based
on Performance
Monitoring Measure(s)
Forecasting
Steps
Data Collection

Data Analysis Track Forecast

Performance (Conditions
May and Often Do
Model Selection Change)

Monitoring
Time Series
Models
• Short Term

• Naïve
• Simple Moving
Average
• Weighted Moving
Average
• Exponential Smoothing
Forecasting Example
• L&F Bakery has been forecasting by “gut feel.”
They would like to use a formal(i.e., quantitative)
forecasting technique.
Forecasting Methods - Naïve
• Forecast for July =
Actual for June

• Ft+1 = At

• FJul = AJun = 600

• Forecast Very Sensitive

to Demand Changes;
Good for stable
demand
Forecasting Methods - Naïve

=C
4
=C
5
 Forecasting Methods – Moving
•Avg
Forecast for July =
Average
of June, May, and April

• Ft+1 = (At+At-1+…)/n

• FJul = (600+500+400)/3 =
500

• Values Equally
Weighted; Good for
stable demand;
Sensitive to
fluctuation; Lags
Forecasting Methods – Moving
Avg

=AVERAGE(C4:C6
)
=
AVERAGE(C5:C7)
 Ft1 = At + At-1 + A t-2 +... + A t-n1
n

Simple Moving Average

You’re manager in Amazon’s electronics
department. You want to forecast ipod
sales for months 4-6 using a 3-period
moving average.
Sale Moving
Month s Average
1 (000
4 NA
(n=3)
2 )6 NA
3 5 NA
4 ? (4+6+5)/3=5
5 ?
6 ?
What if ipod sales were
actually 3 in month 4

Sale Moving
Month s Average
(000 (n=3)
)
1 4 NA
2 6 NA
3 5 NA
4 3? 5
Forecast for Month
5?
Sale Moving
Mont s Average
h1 4
(000 NA
(n=3)
2 )6 NA
3 5 NA
4 3 5
5 ? (6+5+3)/
6 ? 3=4.667
Actual Demand for Month
5=7
Sale Moving
Month s Average
(000 (n=3)
)
1 4 NA
2 6 NA
3 5 NA
4 3 5
Forecast for Month
6?
Sale Moving
Mont s Average
h1 4
(000 NA
(n=3)
2 )6 NA
3 5 NA
4 3 5
5 7 4.667
6 ? (5+3+7)/3=5
Weighted Moving Average
Method
• Used when trend is present
• Older data usually less important

• Weights based on intuition

• Often lay between 0 & 1, & sum
to 1.0

• Equation

Σ(Weight for period n) (Demand in

WMA period n)
= ΣWeigh
ts
Weighted Moving Average: 3/6,
2/6, 1/6
Month Sale Weighte
s d
(00 Moving
0) Average
1 4 NA
2 6 NA
3 5 NA
4 ? 31/6 =
5.167
5 ?
Ft1 = w1At + w 2A t-1 + w 3 A t-2 +...+ w n A t-n1
6 ?
Weighted Moving Average: 3/6,
2/6, 1/6
Month Sale Weighte
s d
(00 Moving
0) Average
1 4 NA
2 6 NA
3 5 NA
4 3 31/6 =
5.167
5 7 25/6 =
4.167
Ft1 = w1At + w 2A t-1 + w 3 A t-2 +...+ w n A t-n1
Exponential
Smoothing
•Assumes the most recent observations have
the highest predictive value
• gives more weight to recent time periods

Ft+1 = Ft + (At - Ft)

et

Ft+1 = Forecast value for Need

time t+1 At = Actual value initial
forecast
at time t
Ft to start.
 = Smoothing constant
 4
9

Exponential Smoothing Equations

Ft+1 = Ft + (At - Ft)

• Premise--The most recent observations
might
have the highest predictive value
• Therefore,we should give more weight
to the more recent time periods when
forecasting
Exponential Smoothing –
Example 1
Ft+1 = Ft + (At - Ft)
i Ai
Week Demand
1 820 Given the weekly demand
2 775 data what are the exponential
3 680 smoothing forecasts for
4 655 periods 2-10 using =0.10?
5 750
6 802 Assume F1=D1
7 798
8 689
9 775
10
3a. Exponential Smoothing –
Example 1
Ft+1 = Ft + (At - Ft)
i Ai Fi
Week Demand = 0.1 0.6
1 820 820.00 82
2 775 820.0 0.0
F2 =
0 F1+ (A1– 0 =820+(820–
3 680 F1) 820)
4 655 815.5 82 =82
5 750 0 0.00
801.9 0
6 802
5 79
7 798
787.2 3.0
8 689 0
6
9 775 72
783.5
10 5.2
3
3a. Exponential Smoothing –
Example 1
Ft+1 = Ft + (At - Ft)
i Ai Fi
Week Demand = 0.1 0.6
1 820 820.00 82
2 775 820.00 0.0
3 680 815.5
F3 = F2+ (A2–
0
=820+(775–
4 655 F2)0 82 820)
5 750 801.9 0.0=815.
5
5 0
6 802
787.2 79
7 798
6 3.0
8 689 0
783.5
9 775 72
3
10 5.2
785.3
3a. Exponential Smoothing –
Example 1
Ft+1 = Ft + (At - Ft)
i Ai Fi
Week Demand = 0.1 0.6
1 820 820.00 82
2 775 820.00 0.0
3 680 815.50 0
4 655 801.9 82
5 750 5 0.0
787.2 0
6 802 This process
6 79 continues
7 798
783.5 3.0
8 689 through week
3 0
9 775 10
785.3 72
10 5.2
8
3a. Exponential Smoothing –
Example
1
Ft+1 = Ft + (A t - Ft)
i Ai Fi
Week Demand = 0.1 = 0.6
1 820 820.00 820.00
2 775 820.00 820.00
3 680 815.50 793.00
4 655 801.95 725.20
5 750 787.26 683.08 What if
6 802 783.53 723.23 the
7 798 785.38 770.49 
constant
8 689 786.64 787.00
equals
9 775 776.88 728.20
0.6
10 776.69 756.28
Exponential
Smoothing
• How to choose α
• depends on the emphasis you want to place on the
most recent data

• Increasing α makes forecast more sensitive to

recent data
• Small alpha  Less importance on recent results
(Good for products with stable demand)

• Large alpha  Recent forecast results more important

(Good for product with varying demands)
Determining Forecast Quality
• How Well Did a Forecast Perform?

• Determine Forecast Error

Error = Actual Demand – Forecasted
Demand
Month Actual Forecast Error
Jan 200 200 0
Feb 300 200 100
Mar 200 230 -30 Average
Apr 400 221 179 Error
121.8
May 500 275 225
Jun 600 343 257
Quantitative Forecasting
Methods Quantitative
Forecasting

Time Regression
Series Models
Models

2. 3. Exponential
1. Naive
Moving Smoothing
Averag
a)simple a)level
e
b)weighted b)trend
c)seasonality
 General Guiding Principles for
1. Forecasting
Forecasts are more accurate for larger groups of
items.

2. Forecasts are more accurate for shorter periods of

time.

3. Every forecast should include an estimate of error.

4. Before applying any forecasting method, the total

system should be understood.

5. Before applying any forecasting method, the

method should
be tested and evaluated.
Summar
y
• What is forecasting
• How does it help a firm?

• What is the difference between potential tools

one may use if the time frame is short term
versus long term?
• Describe the four qualitative forecasting
approaches
• Describe the quantitative forecasting
approaches
• Calculate a simple moving average
• What approach will let you weight more
recent data versus older data?