Going forward, would it be possible to have a configuration option that allows you to take Bayes scores into account when autolearning?
-----Original Message----- From: Matt Kettler [mailto:[EMAIL PROTECTED] Sent: Friday, September 12, 2003 12:49 AM To: Larry Rosenman; [EMAIL PROTECTED] Subject: Re: [SAtalk] What the F****?? At 10:36 PM 9/11/03 -0500, Larry Rosenman wrote: >a SPAM message being learned as HAM? > > >X-Spam-Flag: YES >X-Spam-Checker-Version: SpamAssassin 2.60-rc4 (1.203-2003-08-29-exp) on > lerami.lerctr.org >X-Spam-Level: ***** >X-Spam-Status: Yes, hits=5.5 required=5.0 tests=BAYES_99,HTML_MESSAGE > autolearn=ham version=2.60-rc4 >MIME-Version: 1.0 What's your autolearn threshold set to? Is it greater than 0.2? Remember that autolearning is done based on the score as if there was NO BAYES involved. Thus the bayes_99 goes away. This will VERY significantly reduce the score of this particular message. The only rule left has a very small score: score HTML_MESSAGE 0.160 0.001 0.100 0.100 So, poof, autlearned as ham, even though bayes_99 made it tag by adding another 5.4 points to it. ------------------------------------------------------- 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 ------------------------------------------------------- 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
