The X12 Procedure
- Overview
-
Getting Started
-
SyntaxFunctional SummaryPROC X12 StatementADJUST StatementARIMA StatementAUTOMDL StatementBY StatementCHECK StatementESTIMATE StatementEVENT StatementFORECAST StatementID StatementIDENTIFY StatementINPUT StatementOUTLIER StatementOUTPUT StatementPICKMDL StatementREGRESSION StatementSEATSDECOMP StatementTABLES StatementTRANSFORM StatementUSERDEFINED StatementVAR StatementX11 Statement
-
DetailsData RequirementsMissing ValuesSAS Predefined EventsUser-Defined Regression VariablesCombined Test for the Presence of Identifiable SeasonalityComputationsPICKMDL Model SelectionSEATS DecompositionDisplayed Output, ODS Table Names, and OUTPUT Tablename KeywordsUsing Auxiliary Variables to Subset Output Data SetsODS GraphicsOUT= Data SetSEATSDECOMP OUT= Data SetSpecial Data Sets
-
ExamplesARIMA Model IdentificationModel EstimationSeasonal AdjustmentRegARIMA Automatic Model SelectionAutomatic Outlier DetectionUser-Defined RegressorsMDLINFOIN= and MDLINFOOUT= Data SetsSetting Regression ParametersCreating an MDLINFO= Data Set for Use with the PICKMDL StatementIllustration of ODS GraphicsAUXDATA= Data Set
- References
User-Defined Regression Variables
The X-12-ARIMA method enables you to define regression variables to be included in the regARIMA model. A user-defined regression variable is composed of a value at each time series observation that you provide; the entire variable is implemented as a regressor in the regARIMA model. The regARIMA model is used in the seasonal decomposition process to extend the series prior to X11 decomposition. Because the X-12-ARIMA method does not impute, forecast, nor backcast user-defined regression variables, you must provide a nonmissing value at each observation in the span of the time series to be modeled and also provide a nonmissing value at each observation to be forecast or backcast.
A user-defined regression variable can be included in either the PROC X12 DATA= or AUXDATA= data set. You can supply the values for the user-defined regression variable by one of the following methods:
-
You can include them in an auxiliary data set. The auxiliary data set should have a date variable that corresponds to the date variable in the DATA= data set. The name of the auxiliary data set is specified in the AUXDATA= option in the PROC X12 statement. The name of the date variable that exists in both the DATA= and AUXDATA= data sets is specified in the DATE= option in the PROC X12 statement. The observations in the auxiliary data set must span the entire series plus any forecast and backcast period.
-
You can include them in the DATA= data set. Because the number of observations and the date values are exactly the same for both user-defined regressors and time series values, you need to include forecast and backcast values for user-defined regression variables beyond the span of the time series in one of the following ways:
-
You must specify leading missing values in the series to be seasonally adjusted for backcast periods. You must specify trailing missing values in the series to be seasonally adjusted for forecast periods. You must not use the NOTRIMMISS option in this case. The span of the series to be seasonally adjusted that is implied by trimming the leading and trailing missing values will be less than the span of the date values in the DATA= data set. Using this method, forecast error cannot be computed for the forecast and backcast periods.
-
You can use the SPAN= option in the PROC X12 statement to alter the span of the series to be seasonally adjusted to allow for backcast and forecast periods within the span of the date values in the DATA= data set. Using this method, forecast error can be computed for the forecast and backcast periods.
-
These methods of including user-defined regression variables in the regARMIA model are illustrated in Example 38.6 and Example 38.11.
If missing values for the user-defined regression variable are present within the span of the the time series, including backcast and forecast observations, then an error message is displayed and the time series is not processed. If the span of the user-defined regression variable, or the span after leading and trailing missing values are trimmed, is not sufficient to cover the span of the series to be seasonally adjusted, including any backcasts and forecasts, then an error message is also displayed, and the time series is not processed.