[Spambayes] question regarding training

[Missy]

I have 34 good and 248 spam.  I am not sure if I should do anything else or
just keep things like this.  Any suggestions? 

-----Original Message-----
From: Kenny Pitt [mailto:kennypitt at hotmail.com] 
Sent: Wednesday, August 11, 2004 10:12 AM
To: 'Michael Kimball'; 'Missy'
Cc: spambayes at python.org
Subject: RE: [Spambayes] question regarding training

Michael Kimball wrote:
> Missy wrote:
>> 
>> I mistakenly sent this directly to Tony...my question is, how do I 
>> get it to "train on mistakes"?
> 
> In Configuration | Advanced Configuration | Interface Options set both 
> 'Default training for ham' and 'Default training for spam' to 
> 'discard', and 'Default training for unsure' to 'defer'.

Sorry, I missed it if it was stated previously that this is the
sb_server/pop3proxy version of SpamBayes.  The headers in the original
message seem to indicate that the mail client is Outlook.

If we are, in fact, talking about the Outlook add-in then it is very
difficult to do anything besides "train on mistakes".  Mistake-based
training means that you only train on a message if SpamBayes classified it
as Unsure, of if SpamBayes flat-out got it wrong (i.e. classified a spam as
good or vice versa).  If you simply use the "Delete as Spam" and "Recover
from Spam" buttons in your Unsure folder and on incorrectly classified
messages then you are doing "train on mistakes".

--
Kenny Pitt