What’s New in SAS High-Performance Forecasting 4.1

New ODS plots and plot options are available. You can plot the periodogram for the error series or a combined pediodogram and spectral density estimate plot. You can also generate the prediction error correlation plot matrix.

Auxiliary data set support enables the HPFDIAGNOSE procedure to use other data sets as additional input sources for explanatory variables that are needed during the run of the procedure. Previously all variables that were required during the procedure run had to be physically present in the primary DATA= data set.

The COMBINE statement directs the HPFDIAGNOSE procedure to generate a model combination list for the set of automatic models that are generated from its time series diagnostics. This model combination list is included in the generated model selection list as another candidate forecast for the HPFENGINE procedure to consider in its forecast selection process.

New ODS plots and plot options are available. You can plot the periodogram for the error series or a combined pediodogram and spectral density estimate plot. You can also generate the prediction error correlation plot matrix.

Auxiliary data set support enables the HPFENGINE procedure to use other data sets as additional input sources for explanatory variables that are needed during the run of the procedure. Previously all variables that were required during the procedure run had to be physically present in the primary DATA= data set.

The FORCEBACK option enables you to change the default behavior so that a BACK= region is strictly enforced across all BY groups.

The OUTACCDATA= option directs the HPFENGINE procedure to capture variable information from the run to feed into the HPFTEMPRECON procedure.

The HPFENGINE procedure supports a more general model selection topology. Termed the forecast model selection graph, it remains semantically compatible with the previous model selection list.

The HPFENGINE procedure supports combined models as part of its automated model selection process. User-defined combined model lists are created through the HPFSELECT procedure. The HPFDIAGNOSE procedure can optionally create combined model lists as part of its custom model generation process.

The COMBINE statement directs the HPFSELECT procedure to create a combined model list. Statement options enable you to control different aspects of the forecast combination process for the candidate models identified in the SPEC statements.