On Fri, Jun 06, 2003 at 05:43:00PM +0300, Michael Weber wrote: > Do I need only ham that was marked as spam and spam that > was not marked?
Not only those, but I guess they are important (both when other tests add to the scoring and bayes could help lower the score, and when bayes-based testing made a mistake). > If SA auto-learns, is there any value in sending messages > that were correctly scored back through the process? If they were auto-learned, nope (you'll waste your time as sa-learn skips the ones it has learned already). But if a mail was correctly scored as spam/ham and it wasn't learned, then you can teach the bayes classifier about it and perhaps make bayes decisions better. -- (Mr.) Hannu Liljemark | Appelsiini Finland Oy | http://appelsiini.com ------------------------------------------------------- This SF.NET email is sponsored by: eBay Great deals on office technology -- on eBay now! Click here: http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5 _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
