Thomas> Hmm...I've been wondering something. If you train messages as
Thomas> ham and then later decide that they are spam (or vice versa),
Thomas> does Spambayes get really confused when you flag newer messages
Thomas> or does it slowly shift to unsure and then eventually ham?
Thomas> People's habits _DO_ change over time, for which, I suppose, is
Thomas> a pretty good reason to wipe the database every so often and
Thomas> start over.
If you incorrectly classify a message you will need a few trained the other
way to offset that. I found it's better to find and delete the offending
message. How you do that depends on the particular SpamBayes application
you're using. In some cases you may not have the original messages and be
forced to simply start from scratch. My training databases are stored in
two Unix mbox format files so I just visit them in my favorite mail reader
or text editor and delete them.
Skip
_______________________________________________
[email protected]
http://mail.python.org/mailman/listinfo/spambayes
Check the FAQ before asking: http://spambayes.sf.net/faq.html
