[Spambayes] stupid question
tameyer at ihug.co.nz
Thu Jul 8 02:05:39 CEST 2004
> Whether or not these tokens have any effect on the spam score
> will depend on whether you have trained any messages that
> contain them. At the top of the "show spam clues" you'll see
> the "Significant Tokens", which are the tokens that actually
> contributed to the score given to the message.
One point in addition to what Kenny said:
Unless AVG is only checking a subset of your mail, *every* message will have
this cruft on it (unless you turn it off, or unless it's a virus). So
(assuming a balanced corpus) the clues will stay easily within the 0.4-0.6
range where they are ignored, so they'll have no effect at all (other than a
very small increase in database size).
An "AVG has found a virus" message (I've never seen one of these, but I
suppose they get added, too), would presumably never appear on good mail,
and so quickly end up as strong spam clues, if you train on them.
Please always include the list (spambayes at python.org) in your replies
(reply-all), and please don't send me personal mail about SpamBayes. This
way, you get everyone's help, and avoid a lack of replies when I'm busy.
More information about the Spambayes