[Ifile-discuss] Re: Effect of widely differing volumes on ifile classification

[clemens fischer]

Jack Bertram <address@hidden>:

> Is there any way that ifile could _directly_ age data in its
> database to counter this?

this is getting to be very complex, propably not what ifile stands
for.  but you could try bogofilter if it's only spam.  you will find
lots of bayesian (email) classifiers on sourceforge, but ifile is the
most simple, flexible and easy to use.

for emails, you might need something that does MIME decoding, but this
isn't easy as well.  you cannot simply pipe messages through
something as simple as mimencode, because it breaks headers.  you
would need to separate headers and bodies first.  by the time you're
done with this ...

  clemens