-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On Tuesday 03 February 2004 06:56 am, Pat Traynor wrote:
> Thanks to everyone for suggestions about Bayes, but no one addressed my
> question. �Are there pitfalls to adjusting the scores?
>
> --pat--
I'm sorry I have not read the rest of this thread.
To answer your question though I run my setup that way I have my spam
threshold set at 14 and run most the rule sets that others have posted to
this list. I also have a few of my own and some others that I have found with
searching. I up the scores for html email since in my opinion most of it is
spam and does nothing but clog the internet. Then I up my bayes scores quite
allot. for me this works extreemely well and once the bayes_db is trained
with enough emails false positives are very rare.
I then have maildrop filters setup to hand feed spam and ham to bayes
depending on if they are scored low or high enough. for instance my learn
threshold is set to 35. My delete the spam outright threshold is set to 45.
ham learn is set to -4. In this way I can be sure to catch what I want and
save the marginal stuff for later perusal.
If you would like to see my rule sets and scripts let me know. here is a
taste of my bayes. The auto learn scores here are completely separate from
what I use in maildrop for auto learning
use_bayes 1
auto_learn 1
# Mail which scores outside this range will be fed back into SpamAssassin's
# learning system automatically, to train the Bayesian scanner.
auto_learn_threshold_nonspam 1.0
auto_learn_threshold_spam 20.0
# bayes_ignore_header
# If you receive mail filtered by upstream mail systems, like a
# spam-filtering ISP or mailing list, and that service adds new
# headers (as most of them do), these headers may provide inappropriate
# cues to the Bayesian classifier, allowing it to take a ``short cut''.
# To avoid this, list the headers using this setting. Example:
#
# bayes_ignore_header X-Upstream-Spamfilter
# bayes_ignore_header X-Upstream-SomethingElse
bayes_ignore_header X-s.logic-spamassas-bar
bayes_ignore_header X-s.logic-spamassas
bayes_ignore_header X-SA-Exim
bayes_ignore_header X-Spam-Flags
bayes_ignore_header X-MailScanner-Information
bayes_ignore_header X-MailScanner
bayes_ignore_header X-MailScanner-SpamCheck
bayes_ignore_header X-Spam-Score
bayes_ignore_header X-Scanned-By
bayes_ignore_header X-Sanitizer
bayes_ignore_header ReSent-Date
bayes_ignore_header ReSent-From
bayes_ignore_header ReSent-Message-ID
bayes_ignore_header ReSent-Subject
bayes_ignore_header ReSent-To
bayes_ignore_header Resent-Date
bayes_ignore_header Resent-From
bayes_ignore_header Resent-Message-ID
bayes_ignore_header Resent-Subject
bayes_ignore_header Resent-To
bayes_ignore_header X-Virus-Status
score BAYES_00 -20.0
score BAYES_01 -15.0
score BAYES_10 -10.0
score BAYES_20 -6.0
score BAYES_30 -3.0
score BAYES_40 -1.0
score BAYES_44 -0.5
score BAYES_50 0.5
score BAYES_56 5.5
score BAYES_60 8.5
score BAYES_70 10.5
score BAYES_80 13.0
score BAYES_90 15.0
score BAYES_99 20.0
- --
-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-
Brook Humphrey
Mobile PC Medic, 420 1st, Cheney, WA 99004, 509-235-9107
http://www.webmedic.net, [EMAIL PROTECTED], [EMAIL PROTECTED]
Holiness unto the Lord
-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
iD8DBQFAH77InT1TkA6FgPgRAreNAJoDTO0pjjJTq4/zENZqnW9HAMFF2QCbBoIg
kfEqPFxWorVH9pONQgEBa5k=
=0JX7
-----END PGP SIGNATURE-----
