Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!rochester!udel!gatech!news.cse.psu.edu!uwm.edu!math.ohio-state.edu!jussieu.fr!oleane!tank.news.pipex.net!pipex!news.mathworks.com!uunet!in2.uu.net!news.interpath.net!sas!newshost.unx.sas.com!saswss
From: saswss@hotellng.unx.sas.com (Warren Sarle)
Subject: Re: Input Variable Contribution
Originator: saswss@hotellng.unx.sas.com
Sender: news@unx.sas.com (Noter of Newsworthy Events)
Message-ID: <Dt5pFn.M3H@unx.sas.com>
Date: Mon, 17 Jun 1996 18:16:35 GMT
X-Nntp-Posting-Host: hotellng.unx.sas.com
References:  <31C1A2EE.274E@wvu.edu>
Organization: SAS Institute Inc.
Lines: 20


In article <31C1A2EE.274E@wvu.edu>, Shahab Mohaghegh <shahab@wvu.edu> writes:
|>
|> How would you calculate the contribution of each 
|> input variable (neuron) to the final output?

Even in linear models, there are numerous ways to measure the
contribution of each input. Anybody attempting to answer this
question regarding neural networks should read up on linear
regression first. The best reference I know of for the linear
case is:

  Darlington, R.B. (1968), "Multiple Regression in Psychological 
  Research and Practice," Psychological Bulletin, 69, 161-182.

-- 

Warren S. Sarle       SAS Institute Inc.   The opinions expressed here
saswss@unx.sas.com    SAS Campus Drive     are mine and not necessarily
(919) 677-8000        Cary, NC 27513, USA  those of SAS Institute.
