On Thu, 25 Sep 2003, Matt Kettler wrote: > score FORGED_HOTMAIL_RCVD2 0.051 0 1.884 2.499 > > Notice that in both set0 and set1, it's less that 0.1. Thus the email > scores less that 0.1 when evaluated for autolearning.
Perhaps it would be a good idea to take the difference between the score with bayes and the score without, and skip autolearning if that difference exceeds the required_hits value? ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
