bogofilter's default algorithm

Greg Louis glouis at dynamicro.on.ca
Tue Jan 21 12:10:17 CET 2003


On 20030121 (Tue) at 0934:00 +0100, Boris 'pi' Piwinger wrote:
> David Relson <relson at osagesoftware.com> wrote:
> 
> >>1. does bogofilter -u handle the unknown case (i.e. does nothing)?
> >
> >Correct.  Ham and spam go into the wordlists before and unsure does nothing.
> 
> That makes it unusable to me. How can I make sure, that
> everything is either good or bad?
> 
Set the spam cutoff and nonspam cutoff to the same so there's no unsure
interval.  Probably something between 0.95 and 0.99 will do what you
want.

A note for people who understand the value of unsure: it seems to me
that -u does exactly the wrong thing.  What you want to do is sort the
unsures and train on them; there's little to be gained by training on
only the messages we already got right.

-- 
| G r e g  L o u i s          | gpg public key:      |
|   http://www.bgl.nu/~glouis |   finger greg at bgl.nu |
| Help free our mailboxes. Include                   |
|        http://wecanstopspam.org in your signature. |




More information about the Bogofilter mailing list