From: Troy Settle <[EMAIL PROTECTED]> Date: Mon, 10 Nov 2008 11:30:27 -0500 I received a piece of junkmail this morning: http://home.psknet.com/troy/1.txt In the spam report, I see this: BAYES_00=-2.599 So, I run it through sa-learn with --spam: Learned tokens from 1 message(s) (1 message(s) examined) Then, I re-scan it using spamc, and still I get: BAYES_00=-2.599 What gives? I don't expect the total score to come up much, but the bayes should at least go from a negative number to a positive number... shouldn't it?
The answer Depends on how many tokens bayes is looking at and how spammy those tokens are. You can see what bayes thinks about each token with --debug output. I get BAYES_40 on your message. % wget http://home.psknet.com/troy/1.txt % spamassassin -D --test-mode --debug all,bayes < 1.txt 2>&1 | grep bayes: ... [14389] dbg: bayes: corpus size: nspam = 426975, nham = 53737 [14389] dbg: bayes: token 'Dodge' => 0.999612090680101 [14389] dbg: bayes: token 'sincerely' => 0.999492864983535 [14389] dbg: bayes: token 'decode' => 0.0344385308520192 [14389] dbg: bayes: token 'I'll' => 0.0365668821340277 [14389] dbg: bayes: token 'Perspective' => 0.0404549158471554 ... [14389] dbg: bayes: score = 0.310353325094371 After you learn a message as spam the numbers and raw score should increase somewhat depending on how many times that token has been seen. I get BAYES_60 on the message after learning. % sa-learn --spam < 1.txt % sa-learn --sync % spamassassin -D --test-mode --debug all,bayes < 1.txt 2>&1 | grep bayes: [14618] dbg: bayes: corpus size: nspam = 426990, nham = 53737 ... [14618] dbg: bayes: token 'Dodge' => 0.999615320566195 [14618] dbg: bayes: token 'sincerely' => 0.999498371335505 [14618] dbg: bayes: token 'decode' => 0.0348456512323892 [14618] dbg: bayes: token 'I'll' => 0.0366062570517363 [14618] dbg: bayes: token 'Perspective' => 0.0670493467695761 ... [14618] dbg: bayes: token 'omaha' => 0.958 [14618] dbg: bayes: token 'elsasser' => 0.958 [14618] dbg: bayes: token 'riders' => 0.958 ... [14618] dbg: bayes: score = 0.659988861825694 -jeff
