Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!samsung!sol.ctr.columbia.edu!caen!uwm.edu!mrsvr.UUCP!rupen@gemed.ge.com
From: rupen@gemed (Rupen Sheth x4-6587)
Newsgroups: comp.ai.neural-nets
Subject: Iris plant classification using backprop
Message-ID: <3250@mrsvr.UUCP>
Date: 12 Dec 90 19:05:45 GMT
Sender: news@mrsvr.UUCP
Reply-To: rupen@gemed.ge.com (Rupen Sheth x4-6587)
Organization: GE Medical Systems, Milwaukee,  WI
Lines: 10

I am trying to solve the popular 'Iris Plant Classification' problem using
standard backprop and a network configuration of 4-4-2 (input-hidden-output).
I am using random biases on the hidden and the output layers.

I would like to know what network configurations other people have used with
backprop to solve this problem.  A summary of the configurations and results
would really help.

Thank you.
rupen@gemed.ge.com OR rupen@cvax.cs.uwm.edu