tri-state classification
David Relson
relson at osagesoftware.com
Tue Oct 26 01:41:13 CEST 2004
On Mon, 25 Oct 2004 18:42:09 +0200
Matthias Andree wrote:
> David Relson <relson at osagesoftware.com> writes:
>
> > In the GETTING.STARTED document, particularly the section on
> > "ongoing training", thinking and writing in terms of bogofilter's
> > tri-state abilities seemed especially valuable. Being able to
> > describe messages scores as "Spam", "Ham", and "Unsure" is much
> > clearer than describing them as "X-Bogosity: Yes" and "X-Bogosity:
> > No".
> >
> > I propose that bogofilter's default configuration be changed to use
> > tri-state classification with a conservative ham cutoff of 0.4 and
> > with bogosity tags of "Spam", "Ham", and "Unsure".
> >
> > Let me know if you approve/disapprove of this change.
>
> Go ahead.
An interesting side effect of the small source code changes is that 8 of
the 39 regression tests complain. Several expect "X-Bogosity: Yes/No",
hence need changing to accept "Spam/Ham". Several change from
"X-Bogosity: No, spamicity=0.520000" to "X-Bogosity: Unsure,
spamicity=0.520000". Several expect a return code of 0 (spam) or 1
(ham) and are surprised by 2 (unsure).
'Tis a good reminder for me to put a BIG WARNING in the release notes.
I'll probably release "0.93.0, aka tristate" in a couple of days.
David
More information about the Bogofilter
mailing list