Date:         Tue, 3 Dec 2002 08:56:47 -0700
Reply-To:     William Kossack <kossackw@NJC.ORG>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         William Kossack <kossackw@NJC.ORG>
Organization: Posted via Supernews, http://www.supernews.com
Subject:      Re: scoring a dataset in proc logistic
Content-Type: text/plain; charset=us-ascii

figured it out
the variables in the model statement having missing values eliminate the
calculation of the p value even if the variable is not selected for the model.

William Kossack wrote:

> another thing that puzzles me is that I'm getting a lot of
> missing p values in the first output dataset.  I've looked
> at the values selected by the stepwise and in cases where
> the p value is missing all the parameters have values on
> which to calculate a probability.
>
> William Kossack wrote:
>
> > I'm using stepwise regression in proc logistic.
> >
> > I would like to score a validation data set with the results from the
> > previous logistic
> > run.
> >
> > In my logistic run I'm using
> >
> > proc logistic data=set1 descending outest=parms;
> > class ....;
> > model h_v = ....
> > /selection = stepwise lackfit details;
> > output out=out1 p=p;
> >
> > and for the second run I'm using the same
> >
> > proc logistic data=valid descending inest=parms ;
> > class ....;
> > model v = .... /maxiter=0;
> > output out=out2 p=p;
> >
> > I'm getting
> > NOTE: PROC LOGISTIC is modeling the probability that hospital_visit=1.
> > WARNING: Convergence was not attained in 0 iterations in Step 0. You may
> > want to increase the
> >          maximum number of iterations (MAXITER= option) or change the
> > convergence criteria
> >          (ABSFCONV=, FCONV=, GCONV=, XCONV= options) in the MODEL
> > statement.
> > WARNING: The LOGISTIC procedure continues in spite of the above warning.
> > Results shown are
> >          based on the last maximum likelihood iteration. Validity of the
> > model fit is
> >          questionable.
> > WARNING: Convergence was not attained in 0 iterations in Step 1. You may
> > want to increase the
> >          maximum number of iterations (MAXITER= option) or change the
> > convergence criteria
> >          (ABSFCONV=, FCONV=, GCONV=, XCONV= options) in the MODEL
> > statement.
> > WARNING: The LOGISTIC procedure continues in spite of the above warning.
> > Results shown are
> >          based on the last maximum likelihood iteration. Validity of the
> > model fit is
> >          questionable.
> > ERROR: Missing initial values for some of the parameters in Step 2.
> >
> > Is something is missing or I can score a validation set with the output
> > of stepwise?