10-10-2012 11:19 AM
There are numerous types of forecasting models that a skilled analyst can use. For automatic forecasting of large numbers of time series, only the most robust models should be used.
The candidate models that are used in SAS Forecast Server are considered the most robust in the forecasting literature and these models have proven their effectiveness over time. For more details please have a look at: "Large-Scale Automatic Forecasting Using Inputs and Calendar Events" available here: http://www.sas.com/reg/wp/corp/3478
When using SAS Forecast Studio you will have access to a stored process which also runs a X12 ARIMA model - you can extend these existing stored processes with your own forecasting model at any point in time.
When using HPF code instead of SAS Forecast Studio you can also extend the model library with any type of model you have in mind.
Please refer to my paper: Extending SAS® High-Performance Forecasting Using User-Defined Models and External Forecasts