trinary measures

[Matt Garretson]

Tom Allison wrote:
> I am using the trinary measurement of Yes/No/Unsure with the default 
> cutoff of 0.10 and 0.95.
> What experience do people have with using this as a means of deleting 
> Yes results from the mail server without looking at them.  Is this 99% 


Well it all depends upon how well-trained your database is, and
what rate of false positives you are getting.  So no one besides
you can know the answer to this.   Also, you might consider
spooling at least a day or two's worth of spams in a temporary
location so you can keep an eye on bogofilter's performance, and
can restore an important false positive in a pinch.

At my site i have been using tri-state with a spam_cutoff of .65
and have yet to see what i would consider a false positive.  Just
in case, though, i keep the most recent 2 weeks' spams in a file,
and glance over the low-scoring messages once in a while if i feel
like it.

-Matt