Theo Van Dinter wrote: > There's been discussion about having to have both the original and > recomputed score over/under the spam/ham autolearn score before it'll > actually autolearn, but we haven't really done anything with that yet.
I think that would be a good modification. Right now there are the various problems (of which you are aware) which cause the autolearning to have undesireable artifacts. It is almost like a positive feedback making it unstable. Not autolearning when the current bayes is producing a conflicting result with the standard rules would remove that feedback and hopefully make the algorithm more stable. Bob
pgp00000.pgp
Description: PGP signature