[Spambayes] Problem with POP3 Proxy: [PROBLEM SUMMARY]

[Kenny Pitt]

The numbers represent two different things. The 459 ham/3927 spam is the
number of messages that you have trained, probably because SpamBayes didn't
classify them correctly. The 30976 messages is the number of messages that
SpamBayes has analyzed and classified, which should the same as the total
number of messages that you have received. Does that seem like a reasonable
total number based on the number of messages that you receive?
 
For the two sets of numbers to match, it would mean one of two things: you
are using a train-on-everything strategy (not generally recommended,
especially when you receive so much more spam than good mail), or SpamBayes
got every message that you received wrong (which wouldn't make it a very
good spam filter <wink>).
 
I also notice that your ratio of trained spam to trained ham is tending
toward a potentially dangerous imbalance, probably as a result of the fact
that such a high percentage of your received mail is spam. You may want to
read more about training strategies on the SpamBayes wiki:
 
http://entrian.com/sbwiki/TrainingIdeas
 
-- 
Kenny Pitt
 


  _____  

From: spambayes-bounces at python.org [mailto:spambayes-bounces at python.org] On
Behalf Of Josh Blumenthal
Sent: Thursday, December 16, 2004 8:00 PM
To: spambayes at python.org
Subject: [Spambayes] Problem with POP3 Proxy: [PROBLEM SUMMARY]


I am using SpamBayes POP3 Proxy Version 1.0 (July 2004) (binary), with
version 2.3.3 (#51, Dec 18 2003, 20:22:39) [MSC v.1200 32 bit (Intel)] of
Python; my operating system is Windows 5.0.2195.2 (Service Pack 4).  I have
trained 459 ham and 3927 spam.
 
The problem I am having is that my statistics read as follows:  SpamBayes
has processed 30976 messages - 842 (3%) good, 28721 (93%) spam and 1413 (4%)
unsure.
 
The 30976 msgs seems unrealistic and does not match the # above. A friend is
having the same problem. Suggestions? 
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/spambayes/attachments/20041217/a561094a/attachment.html