Am 16.03.2016 um 03:33 schrieb Ted Mittelstaedt:
On 3/15/2016 6:26 PM, John Hardin wrote:
On Tue, 15 Mar 2016, Ted Mittelstaedt wrote:

we have scripts checking any samples against current bayes
classification and ignore them if they already have BAYES_99,

Is this even necessary? I thought the learner automatically
rejected everything already tagged.

Already *learned*. There's nothing preventing you from learning messages
that scored BAYES_999 (or BAYES_00).


How exactly would it be a bad thing to learn as spam, spam that had this
score? (spam that had been verified, by hand, to be spam - or spam that
arrived at a honeypot address where it would be impossible for it to be
legit)

How would it be a bad thing to learn a piece of spam that had already
been caught by another rule and tagged as spam?

most spam also contains parts which exists in ham too in the hope of bypass filters - the bad thing that happened here after some time was that this hammy tokens got too spammy with no benefit when the message already was classified with BAYES_999 and ham-messages started to lose their BAYES_00 - BAYES_40 classification


Attachment: signature.asc
Description: OpenPGP digital signature

Reply via email to