Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!rochester!cornell!travelers.mail.cornell.edu!news.kei.com!news.mathworks.com!newshost.marcam.com!uunet!cs.utexas.edu!howland.reston.ans.net!math.ohio-state.edu!scipio.cyberstore.ca!nntp.cs.ubc.ca!news.UVic.CA!butterfly!nmicklew
From: nmicklew@butterfly.UVic.CA (Nancy  Micklewright)
Subject: Re: What do you mean by "epoch"? (abused term ?)
Message-ID: <1995Feb28.184417.7920@sol.UVic.CA>
Sender: news@sol.UVic.CA
Nntp-Posting-Host: butterfly.uvic.ca
Organization: University of Victoria, Victoria, BC, Canada
References: <D4FHw9.4qG@unx.sas.com> <3ih0n0$dbb@cantaloupe.srv.cs.cmu.edu> <3iim17$p6u@vixen.cso.uiuc.edu>
Date: Tue, 28 Feb 95 18:44:17 GMT
Lines: 17

Apologies, but I'm still unclear about "epoch" (but soon as I
get an answer, I'm gonna write an article!! (with thanx to
S.F.'s comic relief)).  How about an example to nail this down
for me.
I have 200 cases of input data.  Each case (row) has 10
independent variables, and 1 dependent variable. My proposed
net has: 10 input nodes & a bias node, 1 hidden layer with
6 nodes, and an output layer with one node.
I set my software to update after 16 epochs.
Does that mean I process all 200 cases 16 times, then update,
or will I be processing 16 cases, then updating?  Or am I
off my course and sitting on a water spout?
Thanx.

Steve Berer
nmicklew@nero.uvic.ca

