On Mon, Dec 01, 2003 at 12:55:16AM -0500, Aaron Young wrote: > Has anyone experimented with applying a penalty for small emails? With > bayes or without I always seem to get a mail or two in my inbox. They're > all pretty small, under 1k for the message body. SA finds spammy stuff in > them but it's never enough to go above 5.0. They're anywhere from 3.8 to > 4.9. > > My thoughts were to set a low threshold and a high threshold for > characters. Say 100 and 1000. Then set a score for the rule, such as > 1.0. Email bodies with the number of chars in the low threshold or less > get 100% of the score up to the high threshold which gets 0%. > > Has anyone tried something like this?
I suspect it *really* depends on the sort of mail you typically get. I have a weird habit of mailing myself TODO tasks, reminders, and URLs that I should look at. Those message are very small. To a simple-minded filter, they might look like possible spam. My suggestion is to try it on your corpus of mail and see how it performs. Jeremy -- Jeremy D. Zawodny | Perl, Web, MySQL, Linux Magazine, Yahoo! <[EMAIL PROTECTED]> | http://jeremy.zawodny.com/ ------------------------------------------------------- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
