Not necessarily in more nonspam than spam, but in a significant amount of nonspam. Raising the scores on these (and I tried with each) will severely increase the rate of false positives.
C On Sun, 2002-03-03 at 18:22, Matthew Cline wrote: > On Sunday 03 March 2002 05:58 pm, Craig R Hughes wrote: > > I just pushed out the new scores (and a bugfix or two) as 2.11 > > > > The new scores are done by constraining the GA more, using Michael Moncur's > > submitted scores as a starting point, and then hand-tweaking the output > > where basically any -ve scores that came out but which only existed in the > > corpus as spam (or in the nonspam corpus as highly dubious nonspam) were > > reset to something small and +ve > > Interesting negative scores: > > score INCREASE_SALES -1.273 > score CASHCASHCASH -0.839 > score OPPORTUNITY -0.651 > score DEAR_SOMEBODY -0.492 > score SUBJ_REMOVE -0.471 > score ONCE_IN_LIFETIME -0.405 > score DEAR_FRIEND -0.242 > score MAILTO_LINK -0.226 > score PROFITS -0.162 > score FOR_JUST_SOME_AMT -0.148 > score TO_MALFORMED -0.050 > > So did these things appear in more non-spam than spam? For most of these > rules, I've never seen them appear anywhere but in spam. > > -- > Visit http://dmoz.org, the world's | Give a man a match, and he'll be warm > largest human edited web directory. | for a minute, but set him on fire, and > | he'll be warm for the rest of his life. > [EMAIL PROTECTED] ICQ: 132152059 | > > _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
