Also with the auto learning of spam/ham in Spamassassin, the bayes db will be in a constant state of change, in that more recent spam (score greater than 15 by default) will have precedence and older entries will be expired from the database. This how I understand that it works at least, could be wrong though.
Ryan Moore ---------- Perigee.net Corporation 704-849-8355 (sales) 704-849-8017 (tech) www.perigee.net
Ricardo Kleemann wrote:
Thanks.
I know SA can handle user prefs via SQL, but can it also handle bayes learning from users in SQL?
What I mean is, my email users are in SQL, they are not regular unix users, which means that there's no unix user uniquely mapped for spamc to run as. So I'm stuck with using a generic user, unless SA is capable of running with "sql users" rather than "unix users".
I'm not looking to add user prefs, which I know already is possible with sql. I'm looking for the ability to build bayes databases for "sql users". Is that possible?
Ricardo
----- Original Message Follows -----
Trying to seed the Bayes databases with other people's spam probably isn't a good idea -- the whole idea is that it learns what *your* ham and spam look like. Filling it up with mail from other sources is probably just going to lead to innacurate results.
Ricardo Kleemann wrote:
I've trained my bayes database with about 12,000 spam and 7,000 ham messages, but I was wondering if there are much larger archives available for seeding bayes?
------------------------------------------------------- This SF.Net email sponsored by: Free pre-built ASP.NET sites including Data Reports, E-commerce, Portals, and Forums are available now. Download today and enter to win an XBOX or Visual Studio .NET. http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01 _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk