Re: Recommendations for spam filtering software? [oortish, OTish]

Re: Recommendations for spam filtering software? [oortish, OTish]

[Johan Bockgård]

David Z Maze <dmaze@mit.edu> writes:

> (2) (less directly topical)  Are there other spam filtering tools out
>     there that do a better job?  (Is Naive Bayes just screwed in the
>     presence of somewhat clever spammers?)  It should be pretty easy
>     to change classifiers for spam.el, and since ifile isn't working
>     well for me, I'm not averse to changing.  Options seem to include:
>
>     SpamAssassin (2): Ignore the MIT-mailhub headers and install
>     spamassassin on my own.  No layer-9 issues, filtering does take
>     time though.  (There's also an impression in my mind that
>     spamassassin does "just headers", which doesn't feel reliable to
>     me; more recent versions seem to have checks for spammish bodies,
>     but the indicators aren't particularly informative.)

SpamAssassin works great here. I have had a few false negatives
(mostly in German), and not yet a single false positive.

 It does indeed check the body. A typical example of a mail marked as
 spam by SpamAssassin:

,----
| X-Spam-Status: Yes, hits=17.3 required=4.0 tests=MSG_ID_ADDED_BY_MTA_2,
|   AOL_USERS_LINK,CLICK_BELOW,EXCUSE_3,REMOVE_SUBJ,PORN_4,CLICK_HERE_LINK,
|   CTYPE_JUST_HTML version=2.11
| X-Spam-Flag: YES
| X-Spam-Checker-Version: SpamAssassin 2.11 (devel $Id:
| SpamAssassin.pm,v 1.68 2002/03/04 01:22:24 hughescr Exp $)
| X-Spam-Prev-Content-Type: text/html; charset="iso-8859-1"
| 
| SPAM: -------------------- Start SpamAssassin results ----------------------
| SPAM: This mail is probably spam.  The original message has been altered
| SPAM: so you can recognise or block similar unwanted mail in future.
| SPAM: See http://spamassassin.org/tag/ for more details.
| SPAM: 
| SPAM: Content analysis details:   (17.3 hits, 4 required)
| SPAM: Hit! (1.8 points)  'Message-Id' was added by a relay (2)
| SPAM: Hit! (0.7 points)  BODY: Includes a link for AOL users to click
| SPAM: Hit! (0.9 points)  BODY: Asks you to click below
| SPAM: Hit! (1.0 point)   BODY: Claims you can be removed from the list
| SPAM: Hit! (4.5 points)  BODY: List removal information
| SPAM: Hit! (2.3 points)  BODY: Uses words and phrases which indicate porn (4)
| SPAM: Hit! (1.6 points)  BODY: Tells you to click on a URL
| SPAM: Hit! (4.5 points)  HTML-only mail, with no text version
| SPAM: 
| SPAM: -------------------- End of SpamAssassin results ---------------------
| 
| [body]
`----


/Johan