Hello, Stuart. You wrote 02.01.2005 @ 22:02 in <mid:[EMAIL PROTECTED]> using mailer: The Bat! (v3.0.2.10) Home
SC> My main problems with BayesIt seem to be addressed, but the real SC> issue still seems to exist. This is the act that too much spam is SC> still getting through. As some have mentioned the Ni g erian scam still SC> gets through on a regular basis. Even messages with Vi agra in the SC> Subject line occasionally get through. SC> My question is, does BayesIt look at the subject line for key words SC> and if it does, is there or could there be, more importance placed on SC> words found in the subject line than in the body. It seems that you SC> can have a title that is obvious spam that gets through because of a SC> bunch of irrelevant words in the body. I know about some problems. 2-years experience in implementation of bayesian method give some ideas about how to improve it. However, the "core" of bayesit which parses letters still unchanged during this period. Now, since the method became quite popular, spammers begin to use new technics to cheat it. I am going to change the parsing module - but this is not so easy task to be completed immediately. I am planning to do it in the next version which will be developed after this release. (i've change "Ni ge rian" in your letter because my provider doesn't allow me to send a letter with such word :) ) -- Sincerely, Alexey. Using TB 3.0.2.5 on WinXP Pro SP2 (2600), spelling by ORFO2002 (CSAPI) ..with Kaspersky Antivirus Plugin (ver 3.5 Gold) & antispam filter BayesIt! 0.7.9 (New Year) mailto:[EMAIL PROTECTED] ________________________________________________________ Current beta is 3.0.2.10 | 'Using TBBETA' information: http://www.silverstones.com/thebat/TBUDLInfo.html IMPORTANT: To register as a Beta tester, use this link first - http://www.ritlabs.com/en/partners/testers/
