X13 Procedure

PICKMDL Statement

  • PICKMDL options;

The PICKMDL statement enables you to specify a variety of options for the PICKMDL method. The PICKMDL method uses models that are specified in the MDLINFOIN= data set to choose a regARIMA model. If the MDLINFOIN= option is not specified, then the PICKMDL method chooses a model from the list shown in Table 14. Example 45.9 demonstrates the use of the PICKMDL statement.

The PICKMDL statement cannot be specified when the AUTOMDL statement is also specified. The AUTOMDL and PICKMDL statements each specify different methods of automatic model selection. So only one of these methods can be used to select a model.

For more information about using the US Census Bureau’s PICKMDL method for model selection, see the section PICKMDL Model Selection.

You can specify the following options in the PICKMDL statement:

ARIMAMISS= ARIMASTMT | ZEROORDERS

specifies the method for interpreting missing ARIMA information in a model that is present in the MDLINFOIN= data set. You can specify the following values:

ARIMASTMT

interprets missing information as the model that is specified in the MODEL= option of the ARIMA statement. This option should not be specified if the MDLINFOOUT= data set from a previous X13 procedure call is being used to replicate previous results. However, the (0 0 0)(0 0 0) model is not always the most appropriate model to use as a default when no model has been specified. This option enables you to specify default model orders.

ZEROORDERS

interprets missing information as the (0 0 0)(0 0 0) model. This method is compatible with the output from the MDLINFOOUT= option.

By default, ARIMAMISS=ZEROORDERS.

MDLVAR=variable

specifies the variable in the MDLINFOIN= data set that identifies the models. A model identification variable is not required in the data set if fewer than two models are specified for each series. By default, MDLVAR= _MODEL_.

METHOD= BEST | FIRST

specifies the method for choosing the regARIMA model. You can specify the following values:

BEST

chooses the best model.

FIRST

chooses the first acceptable model.

By default, METHOD=FIRST.

Last updated: June 19, 2025