"The Achilles Heel of the spammers is their message. They can circumvent any other barrier you set up. They have so far, at least. But they have to deliver their message, whatever it is. If we can write software that recognizes their messages, there is no way they can get around that." — Paul Graham
SpamBayes is based on Bayesian filtering, and if you'd like to read the background information, I'll list some links at the end of this article. But my guess is that you don't care to know all the programming details, but just want to get the "body part enlargement" and similar e-mails out of your Inbox so you only have to deal with e-mail you actually want!
Software installation
With that in mind, let's get started. Spam Bayes was written in Python. Again, you probably don't want to deal with installing a Python program so let's do this the easy way. Visit http://starship.python.net/crew/mhammond/spambayes/. Find the "Download SpamBaye-Outlook-Setup-002.exe" link in about the middle of the page, download it, and save it to your hard drive. Close any open applications. Then, simply double click on the file to install it. Accept the default location for installation. At the end of the installation process, there is an opportunity to read the about.html file. This contains helpful instructions for configuring and using the plug-in.
The next time you open Outlook, you'll see a couple of new menu items at the top of the window — SpamBayes has been successfully installed.
Training the software to recognize spam
To begin, create two new folders in Outlook under your Inbox. (I added underlines in front of the names so they would be at the top of my folder list and easy to find.)
I called my folders MaybeSPAM and SPAM. Next, you need to train the SpamBayes. But before you do that, move any spam e-mails you have to the "Spam" folder (or whatever you named it). If you've been diligently deleting spam messages, you probably won't have enough of a sample to use for training. If that is the case, hold off on training. For the next day or so, move all spam messages to the "Spam" folder. Also, make certain that your Inbox (or whichever folder will contain the "good" messages for training purposes) contains only messages you find acceptable.
After you've accumulated a good sampling of each, choose Anti-Spam from the new SpamBayes menu items and then choose Anti-Spam Manager. Click "Train Now" and use the Browse buttons to show the program which folder contains your known good messages and which contains the example spam messages. If this is the first time you've used it, you won't need to check the "Rebuild entire database" checkbox, but you will want to leave the default check in "Score messages after training."
Click "Train Now" and SpamBayes will quickly go through your messages and build a database.
At this point, you can delete messages from the "Spam" folder if you'd like. As new messages arrive, they are classified by SpamBayes as definitely spam or possibly spam and moved to the appropriate folders. When you start using the add-in, you'll want to check those folders periodically to make sure there aren't any false positives. If you do notice any, simply choose Recover from Spam (a new menu item that appears when you click on one of your spam folders) to move it back to your inbox. And, if a new spam message appears in your Inbox, select it and choose Delete As Spam (not simply "Delete") from the menu. The program will learn as you show it your preferences.
Background
For lots of information about Spam, spam filtering research, and SpamBayes, visit Paul Graham's Web site. He even provides links to sites that help spammers to get those unwanted e-mails out to millions of hapless e-mail recipients — in order to help anti-spam programmers figure out how to circumvent the creativity of the spammers.
He mentions two other Bayesian filters for Outlook. Spam Bully is free to try for two weeks then costs $29.95 to register. Spammunition is free.
Other articles of interest on Paul Graham's site:
Will Filters
Kill Spam?
Better Baysian
Filtering
A Plan for Spam
Spam
Resources
Give SpamBayes a try. I'd like to hear what you think of it. me!
© Copyright 2003 by Sound Internet Strategy. All rights reserved