Monday, May 10, 2004

SPAM & POPFile

About six months to a year ago my email traffic spiked to about 500 emails a day with around 98% SPAM. It made my account almost useless, but it was not an account I wanted to lose.

I initially tried filters (a miserable failure), and then gave Mailwasher a try. A good program, but I spent so much time working it, the advantages were almost lost.

In a frenzy of despair I finally came upon Bayesian filtering, and found POPFile to be a popular choice.

Within a day it was over 95% accurate and rising. After a week it was 98% and closing in on 99% accurate. Since December 23, 2003, I have classified 19,479 messages with an accuracy of 99.59%. It dropped a little bit lately as some of the more savvy Spammers are learning tricks to get through the filters. Even so, it is down to no more than a half a dozen a day.

The downside is minor yet worth noting. First, if you are expecting a message from someone who has never emailed you, be a bit paranoid and pay good attention to your mail for the next couple of days. Odds are it will end up in the SPAM folder the first couple of times.

Do not set your email up to empty your trash folder on exit. Every couple of days you will want to go through and check to make sure something didn't make it through. Usually they stand out like a sore thumb so it isn't hard, but you need to do it. If you don't you may lose a half dozen messages in a given month after it is well trained. It isn't bad if it is a newsgroup posting, but if it is your ebilling from your bank, you could have an issue.

All in all I would have to totally revamp my entire email account structure if it were not for POPFile. It still takes a while to download all the messages, but I don't have anywhere near the problem getting through them as I did.

Not only that, but it is free to!

Catch it at http://popfile.sourceforge.net/


No comments: