The X13 Procedure

SEATSDECOMP Statement

(Experimental)

SEATSDECOMP OUT= SAS-data-set <options>;

The SEATSDECOMP statement creates an output data set (named by the OUT= option) that contains the SEATS decomposition series.

The following is an example of a VAR statement and a SEATSDECOMP statement:

   var sales costs;
   seatsdecomp out=SEATS_DECOMP;

The default variable name used in the output data set is the input variable name followed by an underscore and the corresponding table name. Because the B1 series is used as the original input series for the SEATS decomposition, the output data set SEATS_DECOMP from the example will contain the seasonal decomposition variables in the following order:

sales_OS: contains the Table B1 values for the variable sales.
sales_SC: contains the SEATS decomposition seasonal component for the variable sales.
sales_TC: contains the SEATS trend component values for the variable sales.
sales_SA: contains the SEATS seasonally adjusted series for the variable sales.
sales_IC: contains the SEATS irregular component for the variable sales.
costs_OS: contains the Table B1 values for the variable costs.
costs_SC: contains the SEATS decomposition seasonal component for the variable costs.
costs_TC: contains the SEATS trend component values for the variable costs.
costs_SA: contains the SEATS seasonally adjusted series for the variable costs.
costs_IC: contains the SEATS irregular component for the variable costs.

If necessary, the variable name is shortened so that the component name can be added. If you specify the DATE= variable in the PROC X13 statement, then that variable is included in the output data set; otherwise, a variable named _DATE_ is written to the OUT= data set as the date identifier. For further information about the output data set, see SEATSDECOMP OUT= Data Set .

You can specify the following options in the SEATSDECOMP statement:

LEAD=value: specifies the number of periods ahead to forecast for a regARIMA extension of the series. The default is twice the number of periods in a year (8 or 24), and the maximum is 120. In the SEATS computations, the number of backcasts and forecasts are the same, and the minimum number is also dependent on the ARIMA model orders. For more information, see the section SEATS Decomposition. If you specify a LEAD= value that is less than the default, then the number of forecasts specified in the LEAD= option are displayed in the OUT= data set. If the value of the LEAD= option and NBACKCAST= options in the FORECAST statement are less than the required number for SEATS decomposition, then the values of the LEAD= and NBACKCAST= options in the FORECAST statement are increased.
NBACKCAST=value BACKCAST=value NBACK=value: specifies the number of periods to backcast for a regARIMA extension of the series. The default is twice the number of periods in a year (8 or 24), and the maximum is 120. In the SEATS computations, the number of backcasts and forecasts are the same, and the minimum number is also dependent on the ARIMA model orders. For more information, see the section SEATS Decomposition. If you specify a NBACKCAST= value that is less than the default, then the number of backcasts specified in the NBACKCAST= option are displayed in the OUT= data set. If the value of the LEAD= option and NBACKCAST= option specified in the FORECAST statement are less than the required number for SEATS decomposition when SEATSDECOMP is specified, then the value of LEAD= and NBACKCAST= in the FORECAST statement will be increased.
OUT=SAS-data-set: names the data set to contain the SEATS decomposition series: original series, seasonal component, trend component, seasonally adjusted series, irregular component. If the OUT= option is omitted, the data set is named using the default DATAn convention.
YEARSEAS YRSEAS: specifies that two additional variables be added to the OUT= data set: _YEAR_ and _SEASON_. The variable _YEAR_ contains the year of the date that identifies the observation. The variable _SEASON_ contains the month for monthly data, or quarter for quarterly data, of the date that identifies the observation. For monthly data, the value of _SEASON_ is between 1 and 12. For quarterly data, the value of _SEASON_ is between 1 and 4. The _YEAR_ and _SEASON_ variables are useful when you create seasonal plots.