Reconsider the soft drink demand data in Table E4.5 and used in Exercise 4.31.
a. Use the additive seasonal effects model for the first three years to develop a forecasting method for this data. How well does this smoothing procedure work?
b. Make one-step-ahead forecasts of the last 12 months. Determine the forecast errors. How well did your procedure work in forecasting the new data?
c. Compare these forecasts with those found using Winters" multiplicative method in Exercise 4.31.
Exercise 4.31Montgomery eta!. [1990] give four years of data on monthly demand for a soft drink. These data are given in Table E4.5. 4.32
TABLE E4.5 Soft Drink Demand Data
Period
yt
1
143
13
189
25
359
37
332
2
191
14
326
26
264
38
244
3
195
15
289
27
315
39
320
4
225
16
293
28
362
40
437
5
175
17
279
29
414
41
544
6
389
18
552
30
647
42
830
7
454
19
674
31
836
43
1011
8
618
20
827
32
901
44
1081
9
770
21
1000
33
1104
45
1400
10
564
22
502
34
874
46
1123
11
327
23
512
35
683
47
713
12
235
24
300
36
352
48
487
Already registered? Login
Not Account? Sign up
Enter your email address to reset your password
Back to Login? Click here