Hello Andre Wichartz,

on Tue, 24 Feb 2004 18:44:26 +0100 (2004-02-24 18:44:26 in .nl) in the
message with reference <mid:[EMAIL PROTECTED]> you
[AW] wrote (at least in part):

DW>> This is what I've been doing, marking spam as Junk, but shouldn't
DW>> Bayesit start to filter out spam by itself? When does it start
DW>> doing that?

AW> Yes, but in my experience it needs a lot of spam before. At least a few
AW> hundred messages, the more the better. Another thing you could try is
AW> to lower the score needed to move a spam to the junk folder as set in
AW> preferences. You can set it very low, I've set it to 1, the absolute
AW> minimum, and still get no false positives.

Up to the 2.04.04 release BayesIt 0.4gm worked quite fine here, with a
score level of 40.

Starting with 2.04.04, BayesIt 0.4gm SE hardly catches spam anymore. It
has been trained with something like 3000 spam messages, which should be
enough I think. I lowered the score level to 20, without improvement so
far. I'll try your suggested level of 1 for a while.

Also, since 2.04.04, BayesIt is creating files in %SystemDrive%\ (here
D:\) and %Homedrive%%HomePath% (here Z:\), not in any temp-directory.

Was it wise to pack BayesIt 0.4gm SE with TB! on such short notice?
Well, there have been discussions on TBBETA . . .


-- 
Kind regards,
Peter Ouwehand
E: [EMAIL PROTECTED]
- - -
Created the above using
A program which insists to be : The Bat! V2.04.7
An OS which insists to be     : Windows 2000 5.0 Build 2195 Service Pack 4


________________________________________________
Current version is 2.04.04 | 'Using TBUDL' information:
http://www.silverstones.com/thebat/TBUDLInfo.html

Reply via email to