Am 16.02.2015 um 19:33 schrieb Amir Caspi:
Over the last week I've seen a significant uptick in FN spam to my users. We're getting tens of FNs per day per user, whereas a few weeks ago it was just a few FNs per day per user. We're getting BAYES_99/999 on many of these, but no other major markers are hitting (razor, pyzor, dcc, etc.), so maybe we're just the unlucky "early recipients" of most of these spam runs...
unconfirmed herewe maintain a global bayes for spamass-milter and i am doing a hard job in my free-time to collect as much as possible newsletters, registration confirmations, hotel reservations and what not else from some helpful users as well as feed nearly 100% of my personal mail as ham
some of that original messages containing even passwords which i strip out if someone trusting me that enough to provide these samples which is easy since we train from two folders with eml files
given that 24266 messages had BAYES_00 with a total number of 30401 delivered mails in the current month that training strategy seems to work well
[root@mail-gw:~]$ bayes-stats.sh 0.000 0 3 0 non-token data: bayes db version 0.000 0 10470 0 non-token data: nspam 0.000 0 10554 0 non-token data: nham 0.000 0 1368369 0 non-token data: ntokens 0.000 0 993467899 0 non-token data: oldest atime 0.000 0 1424109150 0 non-token data: newest atime0.000 0 1424110103 0 non-token data: last journal sync atime
0.000 0 0 0 non-token data: last expiry atime0.000 0 0 0 non-token data: last expire atime delta 0.000 0 0 0 non-token data: last expire reduction count
insgesamt 35M -rw------- 1 sa-milt sa-milt 2,6M 2015-02-16 19:08 bayes_seen -rw------- 1 sa-milt sa-milt 41M 2015-02-16 19:08 bayes_toks -rw------- 1 sa-milt sa-milt 98 2014-08-21 17:47 user_prefs BAYES_00: 24266 BAYES_05: 396 BAYES_20: 485 BAYES_40: 525 BAYES_50: 3999 BAYES_60: 400 BAYES_80: 293 BAYES_95: 210 BAYES_99: 1629 BAYES_999: 1439 Delivered: 30401 SpamAssassin: 2715
signature.asc
Description: OpenPGP digital signature