[Spambayes] unmarking spam

Jesse Pelton jsp at PKC.com
Thu Feb 21 15:33:02 CET 2008


A few possibilities immediately occur to me; there may be others.
 
The first is simply to keep training on the misclassified messages.
Sooner or later, you'll probably accumulate enough cues that SpamBayes
will classify the messages as you wish.
 
The second is the first on steroids: discard your current training data
and train from scratch. It sounds drastic, but SpamBayes is easy to
train and learns quickly. You should be getting very good results by the
time you've trained on a few dozen messages.
 
A third possibility is to use the SpamBayes Manager to adjust the cutoff
for possible spam. If SpamBayes is assigning the offending messages a
spam probability close to but above the cutoff, you could adjust the
cutoff upward slightly. For example, if SpamBayes typically assigns the
messages an 18% spam probability and the possible spam cutoff is 15%,
you could adjust the cutoff to 20%. The down side to this approach is
that it will affect all messages, and it's possible that you'll get more
spam in your inbox.
 
And finally, there's the option of setting up an Outlook rule to move
messages from that particular sender to some "safe haven" folder (that
is, a folder that SpamBayes is not filtering). SpamBayes attempts to
process messages after any Outlook rules have run, so you should be able
to squirrel these messages away before SpamBayes classifies them. This
is potentially labor-intensive and fragile, and it's somewhat at odds
with the good work that SpamBayes is trying to do for you, so I'd make
this a last resort.
 
If you find a solution that works for you, it would be a kindness to
reply to the list so that your solution will be part of the searchable
mailing list archive for future reference.

________________________________

From: spambayes-bounces at python.org [mailto:spambayes-bounces at python.org]
On Behalf Of Ted Ressler
Sent: Thursday, February 21, 2008 9:04 AM
To: spambayes at python.org
Subject: [Spambayes] unmarking spam


I am using the latest version of spambayes on windows xp.  It works well
except that it continues to mark all email, with or without attachments
from one particular party as suspected spam.  I always click, remove
from spam, but every time something comes in from that party it is
marked as spam.  What can I do?
 
Thanks
Ted
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/spambayes/attachments/20080221/9eca55f9/attachment.htm 


More information about the SpamBayes mailing list