Date: Wed, 12 Sep 2007 06:36:09 -0700
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Shiling Zhang <shiling99@YAHOO.COM>
Subject: Re: logistic regression with given covariance structure
Content-Type: text/plain; charset="us-ascii"
On Sep 11, 4:39 pm, linpan1...@YAHOO.COM (Lin Pan) wrote:
> I have a binary response variable and couple quantitative and qualitative
> variables as predictors. What I want to do is to fit a logistic regression
> model with given covariance structure. I know that if the response
> variable is continuous and normally distributed, I can use proc mixed to
> do it. But I don't know which procedure should I use when the response
> variable is binary. Any help would be very appreciated. thanks.
You can write you own likelihood function in proc nlmixed in stat
pakeage, proc nlp in or pakeage. The proc qlim in ETS pakeage allows
one to specify a certain variance structure, basicly, the variance is
a simple fuction of predictors. It may serves your purpose.