>> I have to get off BayesIT! It learns way to slow. What do people >> think of SpamPal or K9? I have tried both and like both...so, what do >> others think?
IL> Most probable that it been taught badly. IL> It doesn't meter which one program do you use, K9, POPFile, ... all of IL> them have exactly the same logic. IL> The big difference is how you teach your program. IL> My suggestion, delete BayesIT database and teach it again, but do it IL> carefully. 10-20 messages and 90% accuracy of filtering. IL> A few hints, if you have subscribed to any mailing lists, don't mark IL> messages as NOT Junk, put them in white list. IL> White list and black list what you can and let the filter to deal IL> with rest, works for 99.97% in a week. Michael, I dumped BayesIt yesterday afternoon and switched to K9. After three months of training BayesIt with thousands of spam and non-spam messages, it gave me about 10% accuracy. Half a day of K9, and I'm at 90.4% accuracy. It's not a training problem with BayesIt, in my opinion, it's the software. I've used SpamPal for several months on another computer and it does a very nice job. I don't recall it ever making a mistake. Although they work on different filtering principles, I think you'd be happy with either K9 or SpamPal. -- Code 2 :canadaflag: mailto:[EMAIL PROTECTED] The Bat! version 2.12.00 on Windows XP Service Pack 2 In a restaurant window: Don't stand there and be hungry, come in and get fed up. ________________________________________________ Current version is 3.00.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
