[EMAIL PROTECTED] wrote: > Brendon> I've just started using spambayes again after a while away from > Brendon> it. Now, 3 days in, I notice that I've trained on far more > Brendon> spam than ham. (Total emails trained: Spam: *432* Ham: *64) I > Brendon> seem to remember that this was previously my experience in the > Brendon> past. > > Are you training on every message you receive or just the mistakes? Most > I just train on errors and some unsure. At the moment, I'm getting almost no ham in unsure. > people generally only train on the mistakes and unsures. Your ratio is > about 7:1. That's a bit high. I'm maintaining about a 1:1 ratio, currently > That is high relative to the conventional wisdom, but I'm questioning the correctness of that wisdom. > have about 120 each of ham and spam dating back to March (4+ months). (I > Perhaps I receive much more spam than you do! :) > Yes, I believe it was tested pretty thoroughly back in the day. > I couldn't find any reference to such testing and don't remember any from back when Tony and I were writing Spambayes the paper for CEAS. If anybody knows of some research, I'd love to see it.
Perhaps its time to re-evaluate that statement? Brendon. _______________________________________________ SpamBayes@python.org http://mail.python.org/mailman/listinfo/spambayes Check the FAQ before asking: http://spambayes.sf.net/faq.html
