Re: Effect of widely differing volumes on ifile classification

* clemens fischer <[email protected]> [030320 10:57]:
> Jack Bertram <[email protected]>:
> 
> > Recently, the rate, which had been consistent for some time, began
> > to plunge to about 50% and stayed there, until I deleted .idata and
> > rebuilt it from scratch, and it's now classifying better than
> > before.  (Data attached at bottom for completeness)
> 
> how did you do the re-classification and the initial seeding?  you
> can't just feed ifile a big mbox with hundreds of emails, it would
> have to be split before hand.

Well, I actually did it with a modified version of one of Martin Macok's
scripts which I've been using, but it's just as easily done with 

formail -s mbox | ifile -i mbox

assuming mboxes and categories are the same (as they are on my system).
If one mbox could contain more than one category, then I'd have combined
formail with a custom procmail recipe to call ifile with the appropriate
category name, based on the header in the message.

jack