Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!hookup!swrinde!howland.reston.ans.net!newsserver.jvnc.net!nntpserver.pppl.gov!princeton!tucson.princeton.edu!bagalman
From: bagalman@tucson.princeton.edu (Michael D. Bagalman)
Subject: variable elimination
Message-ID: <1995Mar14.153301.19587@Princeton.EDU>
Originator: news@hedgehog.Princeton.EDU
Keywords: neural nets
Sender: mbagalman@attmail.com
Nntp-Posting-Host: tucson.princeton.edu
Organization: Princeton University
Distribution: usa
Date: Tue, 14 Mar 1995 15:33:01 GMT
Lines: 22

I am seeking a little advice, if anyone out there can help me.  I am 
interested in applying neural nets to a problem in which I have more than 
100 predictor/independent variables (inputs) but I know that probably 
only 12 or so are really necessary.

Once I make a neural net, how can I go about determining which inputs are 
unnecessary?  Can I rank them in importance?  Basically I am wishing that 
there is some way, even a primitive unreliable way, to get the equivalent 
of the coefficients and t-tests that you get with linear regression.

Please forgive my ignorance of the topic, one which has probably had 
threads on this newsgroup before.  I would very much appreciate any help, 
either in the form of a direct answer or guidance in the direction of the 
proper journal articles, books, or software packages.

Please either post responses or email to "mbagalman@attmail.com".

Thanks a heap!

Michael


