On Wed, Aug 27, 2003 at 01:59:47PM -0700, Ben Gertzfield wrote: > Just to make sure I wasn't crazy, I backed up my old tokens and seen > files, and re-ran 2.60rc2's sa-learn -D on all my 16,000+ spams. The > end result:
hrm. > Do I just have too many spams? Or is auto-learning somehow messing up > my final result? What DB library are you using? Are there more than the listed count of tokens if you did a "sa-learn --dump"? fyi, 16k is nothing: 0.000 0 2 0 non-token data: bayes db version 0.000 0 155703 0 non-token data: nspam 0.000 0 33877 0 non-token data: nham 0.000 0 334436 0 non-token data: ntokens -- Randomly Generated Tagline: "The adult film industry is like a big family... a big, scary, inbred family." - Hardcore TV
pgp00000.pgp
Description: PGP signature
