On Mar 20, 2004, at 9:47 PM, Brian Dial wrote:
Given the fact that I recieve about 200 times more spam that contains
the habeas headers than actual mail, I've decided toh score habes 0
rendering it useless. However, I cannot seem to reverse the damage it
has done. My guess is that enough of those Fwd: Get All Meds. Fwd:
V|@GRA \ Vali/u/m ( [EMAIL PROTECTED] etc... mails go through and were
autolearned as ham. Now whenever I get one it is scoring BAYES_00 no
matter how often i feed new ones into sa-learn --spam.
You might try
sa-learn --forget on a bunch of recent spam and then sa-learn --spam it?
If that doesn't do it you might need to start bayes over.
$ sa-learn --dump data | grep -i viagra:
0.991 5 0 1078343746 viagra's
0.977 35 49 1079746458 viagra
0.978 2 0 1079594088 VIAGRA
0.958 1 0 1077854779 viagra!
0.958 1 0 1078178284 ViagrAa
0.984 7 5 1079298349 Viagra"..The
0.989 4 0 1079855945 viagra.gif
0.958 1 0 1077496134 Viagral
0.958 1 0 1077858087 viagrav
0.958 1 0 1077578242 viagrast
0.976 36 51 1079594088 Viagra
0.985 3 0 1077240812 viagra-pill.gif
0.049 0 1 1078173704 ViagrAa!
$ sa-learn --dump data | wc -l
166307
$
--
"I don't think the kind of friends I'd have would care."
