There was a very similar feature (the "heuristics" test), but it proved to be too unreliable when it came to mailing list E-mail.I read about this Bayesian filtering/scanning at some other forum as well. Is this something that Declude Junkmail does right now or will do in the (near) future? Would be nice if it were a feature of the scanner on the server in stead of changing all mail client software? ;-)
Although in theory the Bayes Theory should work very well in detecting spam, it does not in reality (for very technical reasons). Using the Bayes Theory for spam testing relies on a number of assumptions that don't hold true -- it's kind of like saying if Sports Team X wins 2 of the first 3 games they play, they have a 66% chance of winning the next game. With the right assumptions, this could be accurate or close to it, but otherwise it just isn't accurate.
-Scott
---
[This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)]
---
This E-mail came from the Declude.JunkMail mailing list. To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type "unsubscribe Declude.JunkMail". The archives can be found
at http://www.mail-archive.com.
