Hello Alexander, Tuesday, October 4, 2005, 4:08:15 PM, you wrote:
ASK> Hello Wolffe & everyone else, ASK> on 04-Okt-2005 at 22:45 you (Wolffe) wrote: >> Is there anyway one can reverse the overtraining? ASK> Just an idea. ASK> You could keep the most recent spam mails, 30 days or so, reset the bayesit ASK> database and then train it anew with the spam mails you kept for spam and ASK> your sent mail folder for ham. You beat me to that. It is something that I started doing when I started beta testing and was having to retrain BayesIt when things did not go well <g> I have kept doing that every month or two. Maybe that is why I have not been having as much touble as others. Also, something that is seldom mentioned, be sure to have penty of 'Ham' for your training as well. Just feeding it spam will make it think everything is spam. You want at least as much ham as spam when you train. -- Best regards, MikeD mailto:[EMAIL PROTECTED] Using The Bat! v3.61.09 Echo (Beta) on Windows XP 5.1 Build 2600 Service Pack 2 ________________________________________________________ Current beta is 3.61.09 (Echo) | 'Using TBBETA' information: http://www.silverstones.com/thebat/TBUDLInfo.html IMPORTANT: To register as a Beta tester, use this link first - http://www.ritlabs.com/en/partners/testers/
