Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news.harvard.edu!news2.near.net!MathWorks.Com!europa.eng.gtefsd.com!howland.reston.ans.net!agate!news.ucdavis.edu!sunnyboy.water.ca.gov!sunnyboy!nsandhu
From: nsandhu@venice.water.ca.gov (Nicky Sandhu)
Subject: Linearly dependent inputs
Message-ID: <NSANDHU.94Oct4165423@grizzly.water.ca.gov>
Sender: news@sunnyboy.water.ca.gov
Organization: Calif. Dept. of Water Resources
Date: Wed, 5 Oct 1994 00:54:23 GMT
Lines: 13

Hi everyone,
	I would like to know if a feedforward neural network where
given inputs which are linearly dependent then 

Q1)  Would that in any way hamper the ability of the neural network
to learn?

Q2)  A related problem is that if a neural network (feed-forward) were
to be used in system identification studies, would it be better to
remove any nearly linearly dependent inputs and replace them with only
one weighted combination of those nearly linearly dependent inputs?
	I would appreciate any advice
-Nicky Sandhu
