Example of Decomposition

A marketing analyst wants to predict sales of a golf driver. The analyst collects previous sales data to predict the sales of the product for the next 3 months.

  1. Open the sample data, GolfDriverSales.MTW.
  2. Choose Stat > Time Series > Decomposition.
  3. In Variable, enter Sales.
  4. In Seasonal length, enter 12.
  5. Under Model Type, choose Additive.
  6. Select Generate forecasts. In Number of forecasts, enter 3.
  7. Click OK.

Interpret the results

The time series decomposition plot shows that the model underpredicts the data at the end of the series. This indicates that Decompostion does not adequately model the trend or the seasonal pattern. The analyst should try Winters' method to determine whether it provides a better fit to the data.

Method

Model typeAdditive Model
DataSales
Length48
NMissing0

Fitted Trend Equation

Yt = 173.06 + 2.111×t

Seasonal Indices

PeriodIndex
1-42.8472
2-32.2639
3-25.4306
4-18.5972
5-1.3056
647.3194
784.1111
830.5278
923.2361
104.1111
11-22.8472
12-46.0139

Accuracy Measures

MAPE7.265
MAD16.621
MSD518.119

Forecasts

PeriodForecast
49233.672
50246.367
51255.312
52264.256
53283.659
54334.396