Date:         Wed, 12 Nov 2003 09:30:38 -0600
Reply-To:     "Peck, Jon" <peck@spss.com>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         "Peck, Jon" <peck@spss.com>
Subject:      Re: estimated coefficients
Comments: To: "Babinec, Tony" <tbabinec@ameritech.net>
Content-Type: text/plain; charset="iso-8859-1"

Introduced in SPSS 12, the Output Management System allows you to save any pivot table (among other things) to a file.  It could be a sav file, XML, text, or html.  Here is a simple example saving the logistic regression coefficients table as a sav file.

oms select tables /if subtypes=['Variables in the Equation'] instances=last
/destination format=sav outfile="e:\temp\coefs.sav".
LOGISTIC REGRESSION  jc3
  /METHOD = ENTER gender educ jobtime
  /CONTRAST (gender)=Indicator
  /CRITERIA = PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .
omsend.

This puts the last coefficients table (there would be two with this syntax) into a sav file.

You can also specify pivots and select columns/dimensions to save.

Regards,
Jon Peck

-----Original Message-----
From: Babinec, Tony
Sent: Wednesday, November 12, 2003 8:53 AM
To: SPSSX-L@LISTSERV.UGA.EDU
Subject: Re: [SPSSX-L] estimated coeficients


So far as I know, Logistic Regression does not have
the capability to save model coefficients.  You can
"by hand" type the model coefficients into a COMPUTE
statement.  Or, to make SPSS do the work, you can
take advantage of the /SELECT keyword on the
Logistic regression command, also found on the dialog
box as Selection Variable.  Use an indicator variable
to score one month's records to one value and the other
month's records to another value, assuming that both of
these sets are in a common file with common variable names.
Then, run Logistic Regression with /SELECT to re-estimate
the model on one set of records and "apply" the fitted
model to the other set.

Anthony Babinec


-----Original Message-----
From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of
Carlos Alberto Auricchio Jr
Sent: Monday, November 10, 2003 9:00 AM
To: SPSSX-L@LISTSERV.UGA.EDU
Subject: estimated coeficients
Importance: High


Dear list,

I´m using the LOGISTIC REGRESSION procedure to estimate the parameters of a
model for one specific month. Then I want to use these estimated coeficients
to compute the estimated probabilities for another month. How could I get
these parameters (not in output)? SPSS save the estimated coefficients in
some hidden variable?

Many thanks in advance!

Carlos Auricchio
Sist. de Informações e Gestão de Autorizações cauricchio@visanet.com.br
tel.: 3457-3435   cel.: 9820-1482