On Thu, Jul 01, 2004 at 10:16:49AM -0400, Jim Maul wrote:
> The are MANY factors that contribute to the decision whether or not to 
> autlearn
> a particular message.  First, there has to be atleast 200 (by default) 
> spam and
> 200 nonspam messages in the bayes database.  Also, bayes autolearning 

Well, not really.  In that case, only manually learning would work
to start a db.  The 200/200 rule is for using Bayes to scan messages.
Autolearning is always enabled unless specifically disabled.

> has to be
> enabled (it is by default) and the SA score of the message has to be within 
> a
> certain range for it to autolearn.  Reading
> http://wiki.apache.org/spamassassin/AutolearningNotWorking would be a good
> start.

:)

-- 
Randomly Generated Tagline:
Any false value is gonna be fairly boring in Perl, mathematicians
 notwithstanding.
              -- Larry Wall in <[EMAIL PROTECTED]>

Attachment: pgpr71tmbSoLy.pgp
Description: PGP signature

Reply via email to