I dug through my archived spam quarantines with the following search:

egrep "<.><\/.>.{0,10}<.><\/.>.{0,10}<.><\/.>" -c /var/backup/spam*

/var/backup/spam-030626:14
/var/backup/spam-030706:14
/var/backup/spam-030713:0
/var/backup/spam-030720:2
/var/backup/spam-030727:0
/var/backup/spam-030810:1
/var/backup/spam-030817:1
/var/backup/spam-030821-1:2
/var/backup/spam-030824:53
/var/backup/spam-030831-2:16
/var/backup/spam-030831-3:6
/var/backup/spam-030907-1:60
/var/backup/spam-030914:61
/var/backup/spam-030921:87
/var/backup/spam-030928:185
/var/backup/spam-031005:164
/var/backup/spam-031012:146
/var/backup/spam-031019:334
/var/backup/spam-031026:661
This query on a mbox of about 500 messages that we tagged high but not spam
(as classified by users) hit 2 times on the above grep.


I ran another grep:
egrep "<[uipb]><\/[uipb]>" -c /var/backup/spam*
/var/backup/spam-030626:577
/var/backup/spam-030706:646
/var/backup/spam-030713:450
/var/backup/spam-030720:772
/var/backup/spam-030727:546
/var/backup/spam-030810:852
/var/backup/spam-030817:1143
/var/backup/spam-030821-1:558
/var/backup/spam-030824:913
/var/backup/spam-030831-2:451
/var/backup/spam-030831-3:99
/var/backup/spam-030907-1:1344
/var/backup/spam-030914:1326
/var/backup/spam-030921:1775
/var/backup/spam-030928:2891
/var/backup/spam-031005:3374
/var/backup/spam-031012:2727
/var/backup/spam-031019:3143
/var/backup/spam-031026:2690
This grep triggered 7 times on the 500 nonspam messages

Each file is one weeks worth of messages (15K-30K) that scored 10+ points.  

These are the ones that I saw being used the most:
<i></i>
<b></b>
<p></p>
<u></u>

Here are some others that I saw:
<C></C>
<T></T>

I'll try testing some rules this week on different variations....

-Matt





-------------------------------------------------------
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

Reply via email to