When I began using GMail I quickly realized that my SPAM automatically ended up in the SPAM folder and was deleted after 30 days. That is very convenient because that leaves a number (unread messages) next to the folder name which I quickly named “SPAM level” – a 30 day history of SPAM. When the FBI or some other organization announced that a mayor bot net had been closed or some spammer jailed I saw the result 30 days later in a drastic decrease in “SPAM level”!

The current rate of false positives (like – once every week – with no regular pattern in sender or content) has left me looking in my SPAM folder a lot! Which is very bad when there is a lot items in there! Having a SPAM folder containing hundreds sometimes thousands of items and trying to find one (ONE!) false positive is literally like looking for a needle in a haystack! It’s simply not worth it to have a SPAM level – so now I clear my SPAM folder every now and then – which means when I check my mail. Fortunately GMail allow me to do that with a single click!

For those who aren’t that informed on the subject I would recommend reading A Plan for Spam by Paul Graham. Wikipedia has a few articles of interest as well. I’d recommend the ones on Bayes’ theorem and Bayesian spam filtering.

Currently I see a lot of areas where this technique can be applied – generally everything evolve around e-mails, RSS-feeds and automatic web-site crawling.

I hope that I’ll have the time in the near future to actually implement something.