SpamAssassin 3.0.1 running under IMail 8.13 (Windoze) is doing really great! I've got about a 1% false negative rate, and way way less fp. Cool!
I know it's critical to sa-(re)learn any FP. I usually sa-learn the false negatives, but sometimes I get lazy. Is that OK? What I mean is, "Is autolearn OK for most use except FPs?" or do I need to train it some more with hams and spams? Magic sez: 0.000 0 3 0 non-token data: bayes db version 0.000 0 9016 0 non-token data: nspam 0.000 0 3139 0 non-token data: nham 0.000 0 173601 0 non-token data: ntokens 0.000 0 1098935201 0 non-token data: oldest atime 0.000 0 1099134563 0 non-token data: newest atime 0.000 0 1099130472 0 non-token data: last journal sync atime 0.000 0 1099108062 0 non-token data: last expiry atime 0.000 0 172800 0 non-token data: last expire atime delta 0.000 0 30072 0 non-token data: last expire reduction count
