Dan Schaefer wrote:
Clip of /etc/mail/spamassassin/local.cf
__________
required_score 7
ifplugin Mail::SpamAssassin::Plugin::AutoLearnThreshold
bayes_auto_learn_threshold_nonspam 0.1
bayes_auto_learn_threshold_spam 10.0
endif
__________
Some messages with a SA score of 10 or higher are auto-learned as spam
and some are not. Any suggestions? What is the default? Perhaps my
bayes_auto_learn_threshold_spam isn't being used. The results seem to
be random. Is there a certain rule that is ignored when determining
the score SA uses for autolearn?
Examples:
score: 11.6 autolearn=no
score: 12.7 autolearn=no
score: 33.9 autolearn=spam
score: 15.9 autolearn=no
score: 19.0 autolearn=no
score: 19.6 autolearn=spam
score: 18.4 autolearn=spam
$ man Mail::SpamAssassin::Plugin::AutoLearnThreshold
----------------------------------------------------------------
Note that certain tests are ignored when determining whether a
message should be trained upon:
* rules with tflags set to ’learn’ (the Bayesian rules)
* rules with tflags set to ’userconf’ (user configuration)
* rules with tflags set to ’noautolearn’
Also note that auto-learning occurs using scores from either
scoreset 0 or 1, depending on what scoreset is used during message
check. It is likely that the message check and auto-learn scores
will be different.
<snip>
Note: SpamAssassin requires at least 3 points from the
header, and 3 points from the body to auto-learn as spam. Therefore,
the minimum working value for this option is 6.
----------------------------------------------------------------
So, Bayes rules and certain other rules are ignored. The score is also
determined using the scores as they would be if Bayes were disabled.
The final score that you see in the email may be significantly different
from the score Bayes is using to determine auto-learn.
--
Bowie
