How to correct Bogofilter wrong decisions

David Relson relson at osagesoftware.com
Fri Jul 27 01:37:13 CEST 2007


On Thu, 26 Jul 2007 22:14:19 +0100
J.R. Seago wrote:

> Yesterday I installed Bogofilter to a new installation of Slackware
> 12, I have canceled the one address which did get spam, so only had
> about four emails that were spam for it to see, today it has decide
> that a perfectly legitimate post to me is spam, and now it has
> decided that emails from the Bogofilter mailing lists are spam. I
> know where the data base is, how do I correct its mistakes, do I just
> keep on changing the 'spam' 'not spam' classifications until it
> learns? -- 
> J.R. Seago
> GNU/Linux Registered User No. #219566 http://counter.li.org/
> () ascii ribbon campaign - against html mail - against microsoft
> attachments /\

Hello J.R.,

Have you read the bogofilter man page and the FAQ?  They have all the
info you need -- notably how to train bogofilter (with multiple ham and
spam messages -- the more the better) and how to correct mistakes.

Your message doesn't specify how many ham and spam you used for
training.  The more training you do, the better bogofilter will be able
to classify messages!

When first using bogofilter, errors in classification are inevitable.
You'll have to use options "-n" and "-s" to expand bogofilter's
knowledge of what's right and what's wrong.

HTH,

David



More information about the Bogofilter mailing list