UNISTAT - the ultimate Excel statistics add-in

9.2.4. Winter’s Multiplicative Seasonal

This adds a mulitiplicative seasonal component to Holt’s Linear method. For a data series xt forecasts are given by:

   Time Series Analysis-Winters Multiplicative Seasonal

Time Series Analysis-Winters Multiplicative Seasonal

where:

·      Time Series Analysis-Winters Multiplicative Seasonal

·      Time Series Analysis-Winters Multiplicative Seasonal

·      Time Series Analysis-Winters Multiplicative Seasonal

·      Time Series Analysis-Winters Multiplicative Seasonal is the level smoothing constant,

·      Time Series Analysis-Winters Multiplicative Seasonal is the trend smoothing constant,

·      Time Series Analysis-Winters Multiplicative Seasonal is the seasonal smoothing constant,

·      s is the season period.

The initial values are based on the complete data series. level values are given by:

   Time Series Analysis-Winters Multiplicative Seasonal

   Time Series Analysis-Winters Multiplicative Seasonal

where:

·      Time Series Analysis-Winters Multiplicative Seasonal is the average level in season i.

The initial seasonal components c1-s, ..., c0 are given by the average value of St, the observed value divided by the expected value if no seasonal component is used. That is:

Time Series Analysis-Winters Multiplicative Seasonal

Time Series Analysis-Winters Multiplicative Seasonal

where:

·      i is the year in which t falls.

·      j is season with in t falls.

Example

Consider the Tasty Cola Sales given in Table 6.4 Bowerman, Bruce L. & Richard T. O’Connell (1987). Open TIMESER and select Statistics 2Forecasting → Winter’s Multiplicative Seasonal and select Cola Sales (C2) as [Variable]. On the second dialogue enter:

·      12 Number of Forecasts.

·      .2 Level Smoothing Constant.

·      .15 Trend Smoothing Constant.

·      .05 Seasonal Smoothing Constant.

·      12 Seasonal Period.

Select the default values on the third dialogue. Select the Summary Table on the Output Options Dialogue to obtain the following results. The table is shortened here for space considerations.

Winter's Multiplicative Seasonal

Level Smoothing Constant =

 0.2000

Trend Smoothing Constant =

 0.1500

Seasonal Smoothing Constant =

 0.0500

Seasonal Period =

 12

Sum of Squares =

 2254.8254

 

Summary Table

Row

Cola Sales

Forecast

Lower 95%

Upper 95%

Level

Trend

Seasonal

1

 189.0000

 193.6418

*

*

 398.0512

 9.2853

 0.4837

2

 229.0000

 238.1693

 214.6797

 261.6590

 404.2000

 8.8148

 0.5838

3

 249.0000

 248.7039

 217.7488

 279.6590

 413.1132

 8.8296

 0.6022

4

 289.0000

 291.9661

 271.1169

 312.8153

 421.5858

 8.9831

 0.6909

5

 260.0000

 252.2613

 233.9958

 270.5268

 433.2106

 9.2472

 0.5866

6

 431.0000

 440.9282

 419.8437

 462.0127

 440.4653

 9.0480

 0.9956

7

 660.0000

 667.1568

 645.5183

 688.7952

 448.5488

 8.9515

 1.4835

8

 777.0000

 774.4063

 754.1714

 794.6412

 457.8068

 8.9822

 1.6929

9

 915.0000

 927.4468

 909.2720

 945.6216

 465.5361

 8.8569

 1.9858

10

 613.0000

 611.8198

 593.9592

 629.6805

 474.5760

 8.8752

 1.2898

11

 485.0000

 487.0187

 470.7199

 503.3175

 483.0504

 8.8351

 1.0072

12

 277.0000

 292.4919

 277.2284

 307.7553

 486.6750

 8.3141

 0.5934

13

 244.0000

 239.4143

 220.1039

 258.7248

 496.8852

 8.5037

 0.4840

14

 296.0000

 295.0392

 275.4274

 314.6510

 505.7181

 8.5366

 0.5839

15

 319.0000

 309.6720

 291.1731

 328.1710

 517.3527

 8.8464

 0.6029

16

 370.0000

 363.5268

 343.7310

 383.3225

 528.0731

 9.0338

 0.6913

17

 313.0000

 315.0635

 295.0703

 335.0568

 536.4034

 8.9634

 0.5864

18

 556.0000

 542.9897

 523.6656

 562.3139

 547.9802

 9.2248

 0.9966

19

 831.0000

 826.6347

 806.6055

 846.6638

 557.7935

 9.2836

 1.4839

20

 960.0000

 960.0143

 940.6599

 979.3688

 567.0755

 9.2835

 1.6929

21

 1152.0000

 1144.5310

 1126.1433

 1162.9188

 577.1112

 9.3587

 1.9863

22

 759.0000

 756.4226

 738.4717

 774.3735

 586.8695

 9.3987

 1.2900

23

 607.0000

 600.5685

 583.2110

 617.9260

 597.5453

 9.5264

 1.0076

24

 371.0000

 360.2125

 342.9294

 377.4955

 610.7077

 9.8900

 0.5941

25

 298.0000

 300.3973

 281.9785

 318.8161

 619.6072

 9.7909

 0.4839

26

 378.0000

 367.4819

 349.3146

 385.6493

 633.0010

 10.1512

 0.5845

27

 373.0000

 387.7550

 368.5888

 406.9211

 638.2576

 9.6617

 0.6020

28

 443.0000

 447.9353

 427.2584

 468.6123

 646.4916

 9.5190

 0.6910

29

 374.0000

 385.4918

 364.6530

 406.3307

 653.3489

 9.0469

 0.5858

30

 660.0000

 660.1998

 638.4242

 681.9754

 662.3557

 9.0409

 0.9967

31

 1004.0000

 996.3279

 975.2618

 1017.3940

 672.4306

 9.1960

 1.4844

32

 1153.0000

 1154.0077

 1133.2125

 1174.8028

 681.5076

 9.1781

 1.6930

33

 1388.0000

 1371.9822

 1351.7913

 1392.1730

 692.2985

 9.4201

 1.9873

34

 904.0000

 905.2261

 885.0484

 925.4038

 701.5284

 9.3915

 1.2899

35

 715.0000

 716.3717

 696.7190

 736.0244

 710.6477

 9.3507

 1.0076

36

 441.0000

 427.7323

 408.5459

 446.9188

 724.4651

 10.0207

 0.5948

37

 

 355.4162

 335.2428

 375.5896

 

 

 

38

 

 435.1913

 414.5975

 455.7850

 

 

 

39

 

 454.2166

 433.1702

 475.2630

 

 

 

40

 

 528.3472

 506.8178

 549.8765

 

 

 

41

 

 453.7447

 431.7040

 475.7853

 

 

 

42

 

 781.9787

 759.4004

 804.5570

 

 

 

43

 

 1179.5346

 1156.3941

 1202.6751

 

 

 

44

 

 1362.2084

 1338.4830

 1385.9339

 

 

 

45

 

 1618.9822

 1594.6506

 1643.3139

 

 

 

46

 

 1063.7802

 1038.8227

 1088.7376

 

 

 

47

 

 841.0292

 815.4278

 866.6306

 

 

 

48

 

 502.4411

 476.1790

 528.7033

 

 

 

 

Time Series Analysis-Winters Multiplicative Seasonal