You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by pinoyskull <pi...@gmail.com> on 2006/11/10 12:00:30 UTC

question about bayes database

will it be ok if i have 1000+ spam learned and only 300+ ham learned, 
will it still be effective?

Re: question about bayes database

Posted by Matthias Haegele <mh...@linuxrocks.dyndns.org>.
pinoyskull schrieb:
> Matthias Haegele wrote:
>> pinoyskull schrieb:
>>> will it be ok if i have 1000+ spam learned and only 300+ ham learned, 
>>> will it still be effective?
>>
>> Dont know. But i think it´s better if you learn *all* spam and ham ...
> that's my problem, spams overwhelmed ham on our server
> 
>> (If your "spam-ham-ratio" is really that bad perhaps you want to use 
>> some "MTA-level" antispam, or blacklists?)
>>
> could you give me an example of a MTA-level antispam, im kinda new to 
> this, thanks

If you use postfix, yes.
general: some "simple" (helo)checks like "dont use my hostname/IP-adress"
a google search for "helo checks (your-mta-name)" should help.

(your-mta-name could be: postfix | qmail | exim ...) ;-).

You could also use blacklists (DUL, abuse etc) for SA (advantage: imho 
scores no "hard" rejection like MTA-blacklist) or your MTA, choose 
blacklist(s) which meet your requirements.

hth
MH

http://www.exit0.us/index.php?pagename=RulesDuJour
http://www.rulesemporium.com/
http://wiki.apache.org/spamassassin/
http://wiki.apache.org/spamassassin/UsingSpamAssassin
http://wiki.apache.org/spamassassin/UsingNetworkTests







Re: question about bayes database

Posted by pinoyskull <pi...@gmail.com>.
Matthias Haegele wrote:
> pinoyskull schrieb:
>> will it be ok if i have 1000+ spam learned and only 300+ ham learned, 
>> will it still be effective?
>
> Dont know. But i think it´s better if you learn *all* spam and ham ...
that's my problem, spams overwhelmed ham on our server

> (If your "spam-ham-ratio" is really that bad perhaps you want to use 
> some "MTA-level" antispam, or blacklists?)
>
could you give me an example of a MTA-level antispam, im kinda new to 
this, thanks
>>
>
> hth
> MH
>
>


Re: question about bayes database

Posted by Matthias Haegele <mh...@linuxrocks.dyndns.org>.
pinoyskull schrieb:
> will it be ok if i have 1000+ spam learned and only 300+ ham learned, 
> will it still be effective?

Dont know. But i think it´s better if you learn *all* spam and ham ...
(If your "spam-ham-ratio" is really that bad perhaps you want to use 
some "MTA-level" antispam, or blacklists?)

> 

hth
MH