Scribd
Upload
68 views

Exponential Smoothing

This document discusses simple exponential smoothing techniques for forecasting. It presents two strategies for initializing the forecasting model: 1) using the first period's actual value, and 2) using a 3-period moving average. Graphs and an error comparison show that initializing with a moving average results in forecasts that more closely track actual demand over time.

Uploaded by

Isabel Del Mar Rivera
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
68 views

Exponential Smoothing

This document discusses simple exponential smoothing techniques for forecasting. It presents two strategies for initializing the forecasting model: 1) using the first period's actual value, and 2) using a 3-period moving average. Graphs and an error comparison show that initializing with a moving average results in forecasts that more closely track actual demand over time.

Uploaded by

Isabel Del Mar Rivera
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 9

EXPONENTIAL

SMOOTHING
IMEN 407
Simple Exponential
Smoothing
Frequently used
Uses weights
Model:

xt , S xt (1 )at 1
xt 1 xt (1 )at 1
Suavizacin exponencial
simple
How to determine a?
Use this formula where N is the number of periods

2

( N 1)
Strategy #1: Using the first period to
initialize
t Dt Forecast
1 52

2 48 52.00

3 36 51.20

4 49 48.16

5 65 48.33

6 54 51.66

7 60 52.13

8 48 53.70

9 51 52.56

10 62 52.25

11 66 54.20

12 62 56.56

Alpha 0.2
Demand and forecast

Demand vs Forecast (SES)


70

60

50

40
Axis Title

30

20

10

0
1 2 3 4 5 6 7 8 9 10 11 12
Strategy #2: Using moving average to
initialize (m=3)

Forecas
t Dt t
1 52
2 48
3 36

4 49 45.33

5 65 46.07

6 54 49.85

7 60 50.68

8 48 52.55

9 51 51.64

10 62 51.51

11 66 53.61

12 62 56.09

Alpha 0.2
Demand and forecast

Demand vs Forecast (SES)


70

60

50

40
Axis Title

30

20

10

0
1 2 3 4 5 6 7 8 9 10 11 12
Comparison

Demand vs Forecast (SES)


70

60

50

40
Axis Title

30

20

10

0
1 2 3 4 5 6 7 8 9 10 11 12
Errors

Error comparison
25

20

15

10

0
1 2 3 4 5 6 7 8 9 10 11
-5

-10

-15

-20

Strat1 Strat2

You might also like