|
> -----Original Message-----
> From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of
> Tanmoy Mukherjee
> Sent: Saturday, January 08, 2011 11:55 AM
> To: SAS-L@LISTSERV.UGA.EDU
> Subject: Re: Logistic Regression Intercept Pr>Chisq
>
> Robin,
>
> Thanks a lot for the explanation. However, I would want to know what you
> mean by not to use Backward Selection as a method for selecting variables.
> I have over 40 variables as dependent variables and I want to select the
> 20-25 best out of them for the model. Therefore based on what I have read
> I believe these are the following options :
>
> 1. Stepwise selection ( Takes a long time to arrive at the best model)
> 2. Selection= Score Best=1 option. ( Takes much faster but does not tell
> much)
> 3. Backwards (Selection= backward RSQ LACKFIT AGGREGATE)
>
> So if this is not the best method I will appreciate if you can tell me
> which method to use to arrive at the best model. How would you select
> variables on your own in the BACKWARD option. I will appreciate if you can
> point me towards any publication that talks about the same.
>
> Regarding your other queries :
>
> 1. The other variables are Statistically significant
> 2. Convergence is established
>
> Thanking you in advance.
Simply search the archives for rants about the problems that selection can cause.
The chorus usually tells you to know your hypothesis, know your field, and to know your model and state it a priori.
In precious few cases might selection perform well. Some posts in the archive will even demonstrate via simulation the dangers.
-Kevin
Kevin Viel, PhD
Senior Research Statistician
Patient Safety & Quality
International College of Robotic Surgery
Saint Joseph's Translational Research Institute
Saint Joseph's Hospital
5671 Peachtree Dunwoody Road, NE, Suite 330
Atlanta, GA 30342
(678) 843-6076: Direct Phone
(678) 843-6153: Facsimile
(404) 558-1364: Mobile
kviel@sjha.org
Confidentiality Notice:
This e-mail, including any attachments is the
property of Catholic Health East and is intended
for the sole use of the intended recipient(s).
It may contain information that is privileged and
confidential. Any unauthorized review, use,
disclosure, or distribution is prohibited. If you are
not the intended recipient, please delete this message, and
reply to the sender regarding the error in a separate email.
|
