[Spambayes] no filterable mail items selected.
Tony Meyer
tameyer at ihug.co.nz
Tue Nov 23 01:21:22 CET 2004
> One particular friend tried to install it and filter
> his pre-sorted emails (SPAM and Good) folders. It will
> not train on them. It keeps saying 2 good, 2 spam, when
> there are actually about 50 in the SPAM folder and 25
> in the Inbox (good) folder.
I suspect (the brand new) FAQ 5.6 applies here:
<http://spambayes.org/faq.html#i-get-an-error-message-no-filterable-messages
-are-selected>
> Being that emails come in 20's and 30's or more at a time,
> its going to take forever to train everything that comes in.
The training that we recommend with the Outlook plug-in is 'mistake based
training'. This means you should only train on messages that end up in the
unsure/possible folder and any good mail in the junk folder or junk in the
good folders. See <http://entrian.com/sbwiki/TrainingIdeas> for more
details.
With this method, you should not need to train on very many of the incoming
messages at all. After even just 5 or so messages of each type, most mail
should be correctly classified. After a couple of days or so SpamBayes
should be classifying with really high accuracy.
=Tony.Meyer
--
Please always include the list (spambayes at python.org) in your replies
(reply-all), and please don't send me personal mail about SpamBayes.
http://www.massey.ac.nz/~tameyer/writing/reply_all.html explains this.
More information about the Spambayes
mailing list