Are there other criteria, though? For example, I have set the threshold at which Auto-Learn is "triggered" for Spam at 7.99. Anything scoring over that is designated to be "auto-learned."
Yet one came through this morning at 12.9, and it did NOT "auto-learn." However, Bayesian probability was something like 0.997, so I thought "maybe at that level it figures it doesn't need to "auto-learn" this one. But when I "hand-feed" it through SA-Learn, it accepts it! So what's the critical difference here? I really want to stop even looking at or in any way dealing with my Spam corpus any more. I want to simply toss anything that is marked as Spam, AFTER it has been "auto-learned". I'm not getting any false-positives any more (and darned few false-negatives). I'm ready to just not worry about it. But I'm afraid that without satisfactory "auto-learn" performance I'm going to veer off course over time. I've yet to see any Spam that has been "auto-learned," after searching diligently for a few days. Ham "auto-learn" seems to work just fine, but not Spam. William L. Polhemus, Jr. P.E. Polhemus Engineering Company Katy, Texas USA -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Martin Radford Sent: Tuesday, October 28, 2003 2:23 PM To: Bill Polhemus Cc: [EMAIL PROTECTED] Subject: Re: [SAtalk] Auto-Learn At Tue Oct 28 18:13:03 2003, Bill Polhemus wrote: > How can I know for sure if Auto-Learn is functioning correctly? Auto-learning is by definition automated, and as such there's a risk that messages will be learned as the wrong type. Learning ham as spam (or vice-versa) can seriously damage the reliability of the bayesian database. To avoid this, there's a significant range in which autolearning is disabled. "autolearn=no" indicates that this particular message was not learned from, not that autolearning as a whole is switched off. ------------------------------------------------------- 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
