home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math.stat
- Path: sparky!uunet!news.claremont.edu!ucivax!ucla-cs!ucla-mic!agsm!iwelch
- From: iwelch@agsm.ucla.edu (Ivo Welch)
- Subject: SAS out-of-sample forecasting routine
- Message-ID: <1993Jan12.111538.13548@mic.ucla.edu>
- Nntp-Posting-Host: risc.agsm.ucla.edu
- Organization: UCLA, Anderson Graduate School Of Management
- Date: 12 Jan 93 11:15:38 PST
- Lines: 17
-
-
- My ultimate goal is to compare different-type regressions on panel data
- sets; some regressions have firm-specific intercepts, others are more
- complex. I would like to forecast with this panel regression on a
- hold-out sample, and produce an MSE. Is there a SAS routine that can
- use a previous model statement and forecast out of sample?
-
- (Note that some variables may be perfectly collinear.) It would be nice
- if SAS could automatically match variables from a "same type" data set,
- especially if there are perfectly collinear variables, which would of
- course disturb the coefficient estimates but not the forecast.
-
- I realize that this might be writeable with some intricate data
- statements. I was hoping this was common enough a request that someone
- had already done this for public consumption.
-
- /ivo welch
-