Date:         Sat, 26 Sep 2009 11:25:21 -0400
Reply-To:     Wensui Liu <liuwensui@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Wensui Liu <liuwensui@GMAIL.COM>
Subject:      Re: Variable reduction method
Comments: To: Peter Flom <peterflomconsulting@mindspring.com>
In-Reply-To:  <9147711.1253961609924.JavaMail.root@elwamui-norfolk.atl.sa.earthlink.net>
Content-Type: text/plain; charset=ISO-8859-1

Peter,
thanks for your insight.
per my limited knowledge of statistical knowledge, there are multiple
implementation of lasso and lar, including the ones you mentioned
applicable to logit model. however, the lasso and lar implemented in
glmselect procedure is for "general linear model" but not "generalized
linear model". therefore, glmselect procedure won't help any in the
logit model PERIOD.



On Sat, Sep 26, 2009 at 6:40 AM, Peter Flom
<peterflomconsulting@mindspring.com> wrote:
> Wensui Liu <liuwensui@gmail.com> wrote
>>
>>Peter,
>>Thanks for sharing your paper again.
>>The paper in the original post is about the variable selection in a
>>multinomial logit. Even I knew that glmselect procedure couldn't be
>>applicable to such a logit model. an expert like you should certainly
>>realize it without reading too much into the paper.
>>
>
> Actually, LASSO and LAR have been applied to multinomial logit, with good results
>
> see Krishnapuram, Carin,Figueireda and Hartemik, SparseMultinomial Logistic Regression
>
> They have also been applied to Cox models and binomial logistic models
>
> Peter
>
> Peter L. Flom, PhD
> Statistical Consultant
> Website: www DOT peterflomconsulting DOT com
> Writing; http://www.associatedcontent.com/user/582880/peter_flom.html
> Twitter:   @peterflom
>



--
==============================
WenSui Liu
Blog   : statcompute.spaces.live.com
Tough Times Never Last. But Tough People Do.  - Robert Schuller
==============================