The HPF Procedure

(View the complete code for this example.)

In the previous example, the HPF procedure was used to automatically select the appropriate forecasting model by using the root mean square error (RMSE) as the default selection criterion. This example illustrates how the HPF procedure can be used to more narrowly specify the possible candidate models. Internet data from the previous example are used for this illustration.

This example forecasts the BOATS variable by using the best seasonal forecasting model (BESTS) that minimizes the mean absolute percent error (MAPE), forecasts the CARS variable by using the best nonseasonal forecasting model (BESTN) that minimizes the mean square error (MSE) using holdout sample analysis, and forecasts the PLANES variable by using the log Winters (additive) method. The following HPF procedure statements forecast each of the transactional data series based on these requirements:

proc hpf data=websites out=nextweek lead=7;
   id time interval=dtday accumulate=total;
   forecast boats   / model=bests criterion=mape;
   forecast cars    / model=bestn criterion=mse holdout=5;
   forecast planes  / model=addwinters transform=log;
run;