Justin Mason said: > To examine how SpamAssassin is making its diagnosis, run with -D -- > this will print out debugging lines for each of the top tokens it > uses to compute the overall result.
hmm, I should go into this in more detail: you want to keep an eye out for lines like: debug: bayes token 'context' => 0.00175165431573987 debug: bayes token 'snip' => 0.00182260025758296 debug: bayes token 'merely' => 0.00189953565131873 debug: bayes token 'forest' => 0.00235725520178676 debug: bayes token 'horror' => 0.00276243104599416 debug: bayes token 'victory' => 0.00302216247655451 debug: bayes token 'filesystem' => 0.00349107827059925 debug: bayes token 'raid' => 0.00354609917849302 the number on the right is key. The more of those you have near 0.0, the lower the bayes score will be -- ie. the more non-spammy. The more near 1.0, the higher it will be. Tokens prefixed with "H:" are mined from the headers. --j. ------------------------------------------------------- This sf.net email is sponsored by: To learn the basics of securing your web site with SSL, click here to get a FREE TRIAL of a Thawte Server Certificate: http://www.gothawte.com/rd522.html _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
