Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!rochester!cornellcs!newsstand.cit.cornell.edu!newstand.syr.edu!news.maxwell.syr.edu!news.bc.net!torn!kwon!watserv3.uwaterloo.ca!novice.uwaterloo.ca!news
From: Stephen Bay <sdbay@watnow.uwaterloo.ca>
Subject: balancing subsets in cross-validation
Sender: news@novice.uwaterloo.ca (Mr. News)
Message-ID: <32FAB0DE.710D@watnow.uwaterloo.ca>
Date: Fri, 7 Feb 1997 04:34:38 GMT
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=us-ascii
Nntp-Posting-Host: watfast.uwaterloo.ca
Mime-Version: 1.0
X-Mailer: Mozilla 3.01 (X11; I; SunOS 5.5 sun4u)
Organization: University of Waterloo
Lines: 12

In cross-validation are the subsets supposed to maintain the same class
proportions as the original dataset? Are there any advantages in not
balancing the subsets?

Thanks

Stephen
-- 
========================================================================
Stephen Bay                 sdbay@watnow.uwaterloo.ca
Systems Design Engineering  http://pami.uwaterloo.ca/~sdbay/stephen.html
University of Waterloo      519-885-1211 x3831
