I cannot speak as an expert on this score, but my understanding is that a
"combination" of such Bayesian databases would likely do more harm than good
because the residual spam corpus that would get through the general SA
filtering would be very specific to the recipient.

That is, combining databases would very likely result in a HIGHER instance
of false positive/false negative, so that you'd actually be WORSE OFF than
if you just let the nature of the Bayesian process take its course with each
separate recipient.

I'd be interested to see other responses, though, by people who know more
than I do about the guts of this algorithm.

William L. Polhemus, Jr. P.E.
Polhemus Engineering Company
Katy, Texas USA

-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Gary
Funck
Sent: Saturday, October 18, 2003 9:33 PM
To: Spamassassin List
Subject: [SAtalk] Easy way to join two Bayes databases?


My personal Bayes database is more up-to-date than a co-worker's. I'd like
to share the database with him,...




-------------------------------------------------------
This SF.net email sponsored by: Enterprise Linux Forum Conference & Expo
The Event For Linux Datacenter Solutions & Strategies in The Enterprise
Linux in the Boardroom; in the Front Office; & in the Server Room
http://www.enterpriselinuxforum.com
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk

Reply via email to