Re[2]: sa-learn

[Robert Menschel <Ro...@Menschel.net>]

Hello LuKreme,

Saturday, August 21, 2004, 7:47:29 AM, you wrote:

>>> But I have had thousands of SPAMS.  Apparently, I've only fed 20 of
>>> them for sa-learn.  I seem to recall that there is a flag to set for
>>> Bayes to pick up on all Spam messages over a certain score.  Have to
>>> go and review again.   

>> from "man Mail::SpamAssassin::Conf"
>>        bayes_auto_learn_threshold_spam n.nn    (default: 12.0)

L> Doesn't that seem unreasonably high?  Doesn't a score of 9.0 have a
L> 100% spam percentage?

Also from Mail::SpamAssassin::Conf:
> required_score n.nn (default: 5)
> Set the score required before a mail is considered spam. n.nn can be an
> integer or a real number. 5.0 is the default setting, and is quite
> aggressive; it would be suitable for a single-user setup, but if you're
> an ISP installing SpamAssassin, you should probably set the default to
> be more conservative, like 8.0 or 10.0.

Following these guidelines required_hits is set at 9.0 for all the
domains I manage, and still we get a handful of false positives each
year.

IMO 12 is probably good for an auto-learn threshold, and I keep mine set
at 15.

Bob Menschel