Date:         Mon, 22 May 2006 05:18:15 -0700
Reply-To:     chandu.isi@GMAIL.COM
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         chandu.isi@GMAIL.COM
Organization: http://groups.google.com
Subject:      Re: How Stepwise Regression works in SAS?
Comments: To: sas-l@uga.edu
In-Reply-To:  <BAY103-F9E7D907FF8E676E5D7917B09A0@phx.gbl>
Content-Type: text/plain; charset="iso-8859-1"

Hi David,

             I can see some of reasons why stepwise regression is not
always better approach( from some posting of SAS-Lers). Thanks for your
info.

            Actually we implemeted some model called " Prescreen Model"
to send some special offers to the customers based on their spending
history of data...something like this.

           So in this case just though that it would better to
understand how STEPWISE REGRESSION works in SAS.

Thanks a lot once again,

Regards,
Chandrasekhara Rao.

David L Cassell wrote:
> chandu.isi@GMAIL.COM wrote:
> >           Is any one knows HOW STEPWISE REGRESSION works in SAS/STAT?
> >Is there any place/URL which will provide this information?
>
> I see that others have pointed you in the right direction.  But let me add
> this warning.
>
> The problem is that stepwise regression DOESN'T work.
>
> Oh, sure, most stat packages have some manner of stepwise regression
> features in there somewhere.  But that doesn't mean it is a good thing
> to do.  Remember that everyone used to store dates with only the last
> two digits for the year, and a 99 for missing.  People finally figured out
> that was a bad thing, too.
>
> If you are thinking about stepwise regression, then I suggest that you
> write back to SAS-L with more details on your problem, and more
> details on why you thought stepwise regression would help you solve
> your problem.  Also tell us how big your problem is, and what version
> of SAS you have.
>
> As things are now, my recommendation would be to look up some of
> my SAS-L posts on stepwise regression in the SAS-L archives, and read
> up on the many problems of stepwise regression.
>
> HTH,
> David
> --
> David L. Cassell
> mathematical statistician
> Design Pathways
> 3115 NW Norwood Pl.
> Corvallis OR 97330
>
> _________________________________________________________________
> Express yourself instantly with MSN Messenger! Download today - it's FREE!
> http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/