|
Training on errors is the regimen that the Outlook plugin
encourages. An "error" is any message that's misclassified: spam that
should be ham, ham that should be spam, or anything classified as unsure. When
you drag a message from, say, the unsure folder to the spam folder, you're
training on an error. Likewise if you select a message and click the "Spam" or
"Not Spam" button. From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of alz Sent: Wednesday, May 24, 2006 11:28 AM To: Jesse Pelton; 'Pam Hook'; [email protected] Subject: Re: [Spambayes] deletion of spam Interesting point! However, how do you train on
errors? From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Jesse Pelton Sent: Wednesday, May 24, 2006 7:58 AM To: Pam Hook; [email protected] Subject: Re: [Spambayes] deletion of spam The short answer: yes. SpamBayes stores the data it uses to
score messages in a database.
The caveat: some people hang on to a collection of messages
(both spam and ham) for retraining. SpamBayes learns quickly enough that I don't
bother with this. If I discard my training and then train on errors (which is
what the Outlook plugin encourages), I generally get good results after just a
few messages. From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Pam Hook Sent: Wednesday, May 24, 2006 8:33 AM To: [email protected] Subject: [Spambayes] deletion of spam I’m running Windows XP Pro. I have read thru the FAQ’s and have not found the answer to my question. I sent you folks an email once before about this question and the only thing I got back was a note asking me to read the FAQ’s.
My question:
My SpamBayes folder is chock full of spam. I would like to permanently delete them, but I’m concerned that if I do, I will need to retrain SpamBayes. Can I delete the spam emails permanently without losing the training files?
Pam
|
_______________________________________________ [email protected] http://mail.python.org/mailman/listinfo/spambayes Check the FAQ before asking: http://spambayes.sf.net/faq.html
