MailScanner, SA and Bayes

[James Gray]

    [ The following text is in the "utf-8" character set. ]
    [ Your display is set for the "US-ASCII" character set.  ]
    [ Some characters may be displayed incorrectly. ]

On Fri, 1 Oct 2004 06:11 pm, Mister PO wrote:
> Hello all,
>
> My config is Postfix 2.1.5, MailScanner 4.32.51, clamAV 0.80rc3, SA
> installed via perl CPAN on a RedHat 9 box.
>
> How can I make sure that Bayes is activated for SA ?
>
> I have checked my /etc/MailScanner/spam.assassin.prefs.conf file and
> Bayes looks OK.
>
> If I run spamassassin
> --prefspath/etc/MailScanner/spam.assassin.prefs.conf - -lint -D, I get
> the following output :

...snipped...
> debug: bayes: found bayes db version 2

This is OK if you're running SA 2.6x.  However, if you're running SA 3.0, 
this should be "db version 3".  If you're running SA 3.0, stop MailScanner, 
and run "sa-learn --sync".  Then restart MailScanner.

> debug: bayes: Not available for scanning, only 1 spam(s) in Bayes DB <
> 200 debug: bayes: 24697 untie-ing

Here's the main problem.  Your bayes database hasn;t seen enough spam yet.  
It requires a minimum 200 ham and 200 spam before it will be used to 
calssify new messages.

You can get a good corpus to "train" bayes with from here:
http://www.spamarchive.org/

..and read the docs about "sa-learn" here:
http://spamassassin.apache.org/full/3.0.x/dist/doc/sa-learn.html

Note that the files from spamarchive.org are in MBOX format, so make sure 
you tell "sa-learn", like this (all on one line, and best to run it as the 
same user MailScanner run's as):

sa-learn -p /opt/MailScanner/etc/spam.assassin.prefs.conf \
--mbox \
--spam corpus.file

Cheers,

James
-- 
Westheimer's Discovery:
        A couple of months in the laboratory can frequently save a
couple of hours in the library.

------------------------ MailScanner list ------------------------
To unsubscribe, email jiscmail at jiscmail.ac.uk with the words:
'leave mailscanner' in the body of the email.
Before posting, read the MAQ (http://www.mailscanner.biz/maq/) and
the archives (http://www.jiscmail.ac.uk/lists/mailscanner.html).

    [ Part 2, Application/PGP-SIGNATURE  196bytes. ]
    [ Unable to print this part. ]