You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Paolo De Marco <pa...@realcomm.it> on 2007/12/20 09:05:33 UTC

Rise up bayes tests

Hi.
Sometimes only bayes tests hit mails, so i recieve mail whit only bayes 
point (for exemple: X-Spam-Status: No, score=3.5 tagged_above=-999 
required=5 tests=[BAYES_99=3.5])
Does anyone raise up the score of bayesan test? Is it safe?

-- 
Paolo De Marco
Real Comm srl
Tel. +39 0434 923134


Re: Rise up bayes tests

Posted by Matus UHLAR - fantomas <uh...@fantomas.sk>.
> > Paolo De Marco schrieb:
> > > Sometimes only bayes tests hit mails, so i recieve mail whit only bayes 
> > > point (for exemple: X-Spam-Status: No, score=3.5 tagged_above=-999 
> > > required=5 tests=[BAYES_99=3.5])
> > > Does anyone raise up the score of bayesan test? Is it safe?

> Matthias Haegele writes:
> > afaik it is not recommended to raise the Bayes Score
> > You could do it but keep in mind if bayes is "misguided" your higher 
> > score hits.
> > (With a well trained bayes it seems reasonable to me)
> > Perhaps you could find some additional rules/network tests ...
> > (sare-rules, razor, dcc, pyzor etc (watch licenses if you could use them)).
> > 
> > On new "few lines text" spam i often get bayes_00 so it is not always 
> > useful.

On 20.12.07 11:53, Justin Mason wrote:
> Actually, that's not quite right -- it's perfectly fine to raise the
> BAYES_99 score, if you feel you've trained it well enough.

IIUC the main problem with higher score is that many people don't know how
to train bayes properly, and then often treat unwanted messages (e.g. from
mailing lists they have subscribed to) as spam.

however the scores 3.5 for BAYES_99 and -2.6 for BAYES_00 are often treated
as too low and I've seen many requests to make them higher (in absolute
value). The problem with adjusting scores is that when someone starts doing
it, many things can happen...

-- 
Matus UHLAR - fantomas, uhlar@fantomas.sk ; http://www.fantomas.sk/
Warning: I wish NOT to receive e-mail advertising to this address.
Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu.
Posli tento mail 100 svojim znamim - nech vidia aky si idiot
Send this email to 100 your friends - let them see what an idiot you are

Re: Rise up bayes tests

Posted by Matthias Haegele <mh...@linuxrocks.dyndns.org>.
Paolo De Marco schrieb:
> Hi.
> Sometimes only bayes tests hit mails, so i recieve mail whit only bayes 
> point (for exemple: X-Spam-Status: No, score=3.5 tagged_above=-999 
> required=5 tests=[BAYES_99=3.5])
> Does anyone raise up the score of bayesan test? Is it safe?

afaik it is not recommended to raise the Bayes Score
You could do it but keep in mind if bayes is "misguided" your higher 
score hits.
(With a well trained bayes it seems reasonable to me)
Perhaps you could find some additional rules/network tests ...
(sare-rules, razor, dcc, pyzor etc (watch licenses if you could use them)).

On new "few lines text" spam i often get bayes_00 so it is not always 
useful.


> 


-- 
Gruesse/Greetings
MH


Dont send mail to: ubecatcher@linuxrocks.dyndns.org
--