On Tue, Dec 02, 2003 at 01:06:55PM -0600, Bob Apthorpe wrote: > > 1. Can anybody tell me how to read the values obtained from sa-learn --dump? > > The results of my dump are below: > > > > 0.000 0 2 0 non-token data: bayes db version > > 0.000 0 49175 0 non-token data: nspam > > You've trained on 49175 pieces of spam...
Yep, and you're running with DB version 2, which means you're using 2.6x. :) > > 0.000 0 0 0 non-token data: nham > > ... but no ham. Learn from at least 200 pieces of ham (sa-learn --ham ...) > so Bayesian scoring kicks in. yep. actually, with 49k spam, you'll want way more than 200 ham to be accurate. Most everything else is self explanatory I think. 0.000 0 125566 0 non-token data: ntokens that's tokens in the db. 0.000 0 1069458380 0 non-token data: oldest atime 0.000 0 1070374969 0 non-token data: newest atime atime (time_t) of the oldest and newest tokens in the db. 0.000 0 0 0 non-token data: last journal sync atime your journal hasn't been synced yet. it'd be an atime value as well. 0.000 0 1070114468 0 non-token data: last expiry atime your last expiry run occured at this atime (Sat Nov 29 09:01:08 2003 EST) 0.000 0 571530 0 non-token data: last expire atime delta 0.000 0 67720 0 non-token data: last expire reduction count "man sa-learn" has more information about how these values are used. it's there to do estimation values during the next expire run. -- Randomly Generated Tagline: "If Linux is really to gain "world domination" making sure that NT's strengths are Linux's strengths as well as making sure that NT's weaknesses are Linux's strengths is required." - Paul Flinders
pgp00000.pgp
Description: PGP signature
