Hello Greg, Sunday, October 10, 2004, 8:21:19 PM, you wrote:
GS> Hello MikeD, GS> Thursday, October 7, 2004, 10:11:36 AM, MikeD (3) wrote: >> Click Add and then point it at the 'bayesit' folder and select the >> 'bayesit.tbp' GS> Done. >> That will get things started. Next you have to start showing it what >> is good email (i.e. 'ham') and what is spam (that is train it). You >> can do that in several different ways. You can just start using TB as >> before. Without anything to go on it *should* treat everything as >> ham. You then tell it what is spam. Eventually it will start to >> figure things out. GS> Ok I can do nothing and let learn on the fly, but I have about 500 GS> spam messages that I would like to use to train. >> If you have saves a lot of spam and ham samples, you can train it on >> those saved letters. That will speed up the training process. GS> How do you do this? Select all the spam messages the way you would select any messages and once you have them all selected, click on Specials and select "Mark as Spam" from that drop down menu. That is all there is to it. However, a word of caution: with only spam (and no "ham") in the training base, you are likely to get a lot of false positives because there will be lots of "spam" indicators from the spam messages you trained it with, but no (or almost no) offsetting "ham" indicators. You would be better served to wait until you have processed some "ham" mail so that the system will have some indication of what a good message will look like. One idea that someone else suggested was to use your "sent mail" for "ham" training. That might work unless you are like me and get a lot of "news letters", "news items" and the like that are nothing like what I send out <g> Hope that helps. -- Best regards, MikeD mailto:[EMAIL PROTECTED] Using The Bat! v2.12.00 on Windows ME 4.90 Build 3000 ________________________________________________ Current version is 3.00.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
