I finally got BayesIt trained and it caught its first bit of spam today. :) A few questions:
1. Is ongoing training necessary only when BayesIt misses spam or mis-identifies non-spam, or is something more required? 2. Is there a point where training BayesIt doesn't noticeably improve accuracy, and in fact slows down processing because of of the size of its dictionaries? If so, where is that point? 3. You guys describe your own accuracy stats in terms of "99.8% for last month", "99.94% for last week" and such. Where are you getting these stats? And for you Ritlabs guys watching: you've *got* to get the help files updated and/or a manual produced. BayesIt would be a lot less intimidating with some instruction. -- Code 2 mailto:[EMAIL PROTECTED] The Bat! version 2.11.02 on Windows XP Service Pack 1 On another Butcher's window: Pleased to meat you. ________________________________________________ Current version is 2.11.02 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
