Bryce's Radio Experiments
Musings on general technology.

Permanent Link Friday, November 01, 2002

POPFile, Part III

The mail parser has been updated to handle Outlook .MSG files.

There's a thread on corpus drifting that covers my thoughts on using positive reinforcement to help POPFile to learn. On the mailing list I am training POPFile on, it has missed 3 of 22 messages today. I'm thinking that POPFile needs about 100 messages in the corpus to get accuracy into the high 90s for mailing lists.

On the spam front, I seem to be in the middle of a drought. POPFile has missed 1 of 5 messages since yesterday.

I've found another bug, POPFile seems to top out at 8 simultaneous connections. I have 10 POP accounts in three of Outlook's "Send/Receive Groups." They have staggered times for checking mail but every so often they all overlap...

11:48:27 AM | Comments: | Topics: bayesian spam 


© Copyright 2003 T Bryce Yehl Click here to send an email to the editor of this weblog.
Last update: 6/29/2003; 10:00:24 PM.
the