Bayes?

[Soeren Gerlach]

> Doe anyone who is running SA2.5x know the answer to the following
> question... When putting a 'false positive' through sa-learn do you
> first need to remove {SPAM?} from the Subject and the X-MailScanner
> headers?

As sa-learn analyzes the whole message (but knows the difference between a
message header, it's body and subject) it is recommended (regarding to the
docs) to use the message before it has passed MS and has been tagged by it
with footer, mail header lines etc. to get "clean" results in the Bayes
database. But I don't know what happens if you feed some 1,000s messages
into the database all with a "Scanned by MailScanner...." footer in them,
my guess is that it will treat just these words a little like spam.

regards,
Soeren Gerlach