On Wed, 13 Aug 2003, Matt Kettler wrote: > At 10:01 AM 8/13/2003 -0400, Allin Cottrell wrote: > > If I don't use sa-learn manually, how is sa supposed to be > >made aware of its degree of non-success (the false negatives)? > > > It can't. > > Autolearning isn't a perfect solution, it just provides a reasonable > mechanism where some learning can occur without manual effort. You should > still be training it using sa-learn whenever practical, especially when it > has false positives and negatives, and very especially when the BAYES_ rule > that matched is on the wrong side of the center line (50).
Thanks for the reply, and I'm quite happy to use sa-learn from time to time. But I'm afraid I'm still in the dark on my original question. I don't see how *any* learning can occur in the absence of explicit feedback from the user on whether sa has classified messages correctly or not. To rephrase my question: how can this feedback be given, other than by using sa-learn? Allin Cottrell. ------------------------------------------------------- This SF.Net email sponsored by: Free pre-built ASP.NET sites including Data Reports, E-commerce, Portals, and Forums are available now. Download today and enter to win an XBOX or Visual Studio .NET. http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01 _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
