By Deane Barker on June 27, 2003
SpamBayes Outlook Addin: Here’s a spam filter for Outlook that runs on Bayesian theory. You give it a folder of good emails (your inbox) and a folder of spam that you’ve collected, then let it analyze both.
From then on, it will use this information to assign a score to each incoming message and either leave it in the Inbox, banish it to a spam folder, or put it in an “unsure” folder where you can review it. You can set the thresholds: for me, a message is for sure good if the “spam score” is less than 15, it’s for sure spam if the score is above 90, and the rest is unsure.
You can manually tag messages as spam, or rescue them as legitimate, and the filter incorporates that action into future classification — essentially you train the filter as to what you consider spam and what you don’t. The longer you use it, the more bulletproof it supposedly becomes and the tighter you can make the thresholds.
It installs beautifully, and seems to work pretty well right out of the box. I let it analzye about 100 good emails and 300 spams (I had to shut SpamKiller off for a few hours to collect them), and it had very few mistakes. I’m training mine now, and I’ll update when I decide if it works or not.