[Spambayes] Spam Prefiltering
Michael C. Neel
neel at mediapulse.com
Sat May 22 17:49:25 EDT 2004
I'd recommend letting all mail goto the inbox, even better if the other
filter is turned off completely. Blacklisting is bad, FAQ and threads on
this about, lol.
Then train on a balanced set of spam/ham (i do 100 ea), and only train the
ones spambayes gets wrong. Check on the database every now and then, and if
the numbers get too far off balance, add some more spam/ham to the training
to balance it out.
SB currently handles an account I get over 400 spams a day, and with the
above message I see maybe a few emails a day in suspects, and occasoinally
my brother emails me an ebay link I have to get out of the spam folder, lol.
----- Original Message -----
From: "David Abrahams" <dave at boost-consulting.com>
To: <spambayes at python.org>
Sent: Saturday, May 22, 2004 5:32 PM
Subject: [Spambayes] Spam Prefiltering
> I just discovered that my sysadmin was pre-filtering my incoming
> mail. Man, it's amazing how much spam I'm getting! The problem is
> that it was going into a mailbox I was using for nightly training:
> DA> Hi Andrey,
> DA> Have you got some kind of spam filtering going on our accounts,
> DA> in addition to what I'm doing personally with SpamBayes? My
> DA> SpamBox mailbox has practically doubled in size over the past
> DA> week, and I don't have any rules that put messages in there --
> DA> I'm only using it for training SpamBayes. Is there some other
> DA> automated system putting messages in there? Since no new
> DA> messages are going into my HamBox, they're getting
> DA> out-of-balance, which throws the training off.
> DA> Thanks,
> DA> Dave
> I'm using several blacklists (spamcop, spamhaus etc.)
> When mail is coming from IP address in these lists, Communigate
> adds suffix 'blacklisted' to the address of recipient. After that
> it's possible to:
> 1) deliver mail to SpamBox (as it going now)
> 2) deliver mail to Inbox
> 3) reject mail
> I can route all mail to your Inbox, if you want it.
> So now the question is, what to do? Clearly I don't want to route
> these into my spam training folder automatically unless I have a way
> to balance them with Ham. Will I get better results from SpamBayes if
> I give it a chance to learn from these messages (i.e. send it to my
> INBOX and let SpamBayes filter them), or should I just discard them?
> Any Advice appreciated.
> Dave Abrahams
> Boost Consulting
> Spambayes at python.org
> Check the FAQ before asking: http://spambayes.sf.net/faq.html
More information about the Spambayes