| Tobias DiPasquale on Wed, 19 Mar 2003 07:10:19 -0500 |
|
On Wed, 2003-03-19 at 02:09, sean finney wrote: > On Tue, Mar 18, 2003 at 11:14:39PM -0500, Jeff Abrahamson wrote: > > words, 141 characters). It's attached, for your amusement. I think > > it's cool to know just how many words you wrote me when I read your > > email. I'm trying to think of more interesting analyses to do. > > if you want to go really crazy with it, how about an analysis for word > frequency? you could keep word counts of this list in some kind of > giant histogram, and then have a program that tries to guess the > sender from the content of his/her message :) What you've described is naive Bayes. Its in use in programs like POPFile and Bogofilter already, bogofilter having built-in hooks into mutt currently. -- Tobias DiPasquale 88FA 30C9 1E63 CFE2 CBD8 37C4 DA1C E2BF 1D26 F036 http://cbcg.net/ Attachment:
signature.asc
|
|