-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Martin Radford writes: >At Fri Oct 17 21:17:54 2003, Ben Wing wrote: >> >> hi. i just upgraded from 2.53 to 2.6 and i'm seeing something >> rather odd about the Bayesian results: nearly every one is almost >> exactly 0%, 50%, or 100%! it's almost as if it's applying an >> extreme rounding function to the actual result. now, these are >> turning out so far to be accurate, but i'm still highly distrustful >> of such "perfect" results. this clustering happened the instant i >> upgraded spam assassin -- in fact, one of the first messages i sent >> after this > >I found this when I first upgraded to one of the pre-releases of 2.60. >The developers said that this was due to changing the method of >calculating the Bayes score. The newer code is much more likely to >cluster around 0, 0.5, and 1. I have seen a few messages outside >those cluster areas, but not too many. I've not seen any FPs, though. If you're seeing FPs, it's strongly indicative of mistakes in the training data -- spam trained as ham or vice-versa, I'm afraid ;) - --j. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.2 (GNU/Linux) Comment: Exmh CVS iD8DBQE/kxMjQTcbUG5Y7woRAgnyAJ9GaPCdey9oNgAT/y2ZiJkahjPuIgCgoxAC vPt8S4fWAKrhfkvq++O4BmI= =JWtb -----END PGP SIGNATURE----- ------------------------------------------------------- This SF.net email sponsored by: Enterprise Linux Forum Conference & Expo The Event For Linux Datacenter Solutions & Strategies in The Enterprise Linux in the Boardroom; in the Front Office; & in the Server Room http://www.enterpriselinuxforum.com _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
