Overview (SEASON command)

SEASON estimates multiplicative or additive seasonal factors for time series by using any specified periodicity. SEASON is an implementation of the Census Method I, otherwise known as the ratio-to-moving-average method. See 1 and 2.

Options

Model Specification. You can specify either a multiplicative or additive model on the MODEL subcommand. You can specify the periodicity of the series on the PERIOD subcommand.

Computation Method. Two methods of computing moving averages are available on the MA subcommand for handling series with even periodicities.

Statistical Output. Specify TSET PRINT=BRIEF to display only the initial seasonal factor estimates. TSET PRINT=DETAILED produces the same output as the default.

New Variables. To evaluate the displayed averages, ratios, factors, adjusted series, trend-cycle, and error components without creating new variables, specify TSET NEWVAR=NONE prior to SEASON. This can result in faster processing time. To add new variables without erasing the values of previous Forecasting-generated variables, specify TSET NEWVAR=ALL. This saves all new variables that are generated during the current session to the active dataset and may require extra processing time. When the default (TSET NEWVAR=CURRENT) is in effect, only variables from the current procedure are saved to the active dataset, and the suffix #n is used to distinguish variables that are generated by different series. TSET MXNEWVAR specifies the maximum number of new variables that can be generated by a procedure. The default is 60. The order in which new variables are added to the active dataset's dictionary is ERR, SAS, SAF, and STC.

Basic Specification

The basic specification is one or more series names.

  • By default, SEASON uses a multiplicative model to compute and display moving averages, ratios, seasonal factors, the seasonally adjusted series, the smoothed trend-cycle components, and the irregular (error) component for each specified series (variable). The default periodicity is the periodicity that is established with TSET PERIOD or DATE.
  • Unless the default on TSET NEWVAR is changed prior to the procedure, SEASON creates four new variables, with the following prefixes, for each specified series (these variables are automatically named, labeled, and added to the active dataset):

    SAF. Seasonal adjustment factors. These values indicate the effect of each period on the level of the series.

    SAS. Seasonally adjusted series. These values are the values that are obtained after removing the seasonal variation of a series.

    STC. Smoothed trend-cycle components. These values show the trend and cyclical behavior that are present in the series.

    ERR. Residual or "error" values. These values are the values that remain after the seasonal, trend, and cycle components have been removed from the series.

Subcommand Order

  • Subcommands can be specified in any order.

Syntax Rules

  • VARIABLES can be specified only once.
  • Other subcommands can be specified more than once, but only the last specification of each one is executed.

Operations

  • The endpoints of the moving averages and ratios are displayed as system-missing in the output.
  • Missing values are not allowed anywhere in the series. (You can use the RMV command to replace missing values, use TSET MISSING=INCLUDE to include user-missing values, and use USE to ignore missing observations at the beginning or end of a series. See RMV and USE for more information.)

If SPLIT FILE processing is on, SEASON ignores USE settings.

Limitations

  • A maximum of one VARIABLES subcommand is allowed. There is no limit on the number of series that are named on the list.
1 Makridakis, S., S. C. Wheelwright, and V. E. McGee. 1983. Forecasting: Methods and applications. New York: John Wiley and Sons.
2 McLaughlin, R. L. 1984. Forecasting techniques for decision making. Rockville, Md.: Control Data Management Institute.