John, if you run the email through Spamassassin -tD <mail.txt |more then you should be able to see which tokens it's finding, and the weight it's giving. Perhaps that particular email contains tokens which have been found in several learnable hams as well. If you save the output at the initial test, and then test again when this happens, you may see that those tokens get weighted differently. -tom
> -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On > Behalf Of Stewart, John > Sent: Monday, November 03, 2003 1:13 PM > To: '[EMAIL PROTECTED]' > Subject: [SAtalk] Problems with bayes "forgetting" in 2.60 > > > > However, what I've noticed is that bayes seems to be > "forgetting". I had a couple of mails I trained it on last > week, and immediately after learning them, it was hitting > BAYES_99. Today one is hitting BAYES_50, and the other is not > hitting any bayes rules, which I take to mean the check_bayes > algorithm is returning somethintg between 0.4999 and 0.5001, > as this is the only area not scored by some amount in the > 23_bayes.cf file. > ------------------------------------------------------- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk