A Better Title

Having EE2008 Day Structure and Algoristmas now… Currently the lecturer is going through the divide and conquer method which is the method supposed to be used for merging and sorting arrays. Which I did go through back when I was still in JC taking computing.

Before this was EE2006 Engineering Mathematics I. We went through this cool thing called Bayes’ Theorem and its use in Probability and Statistics. The cool thing is that Bayes’ Theorem is the exact same thing being used in POPFile, the program I use together with Mozilla Thunderbird to sort out which emails are supposed to be personal, blog comments, spam, from Friendster, NTU Stuff, subscribed groups, and subscribed emails.

POPFile analyses incoming emails for the words contained and assigns a category for the emails. It can be trained by telling it which category an email is supposed to go to if it classifies wrongly, and it will learn and adjust the probability numbers of the words contained in its database in such a way that the next time it encounters the same words, it will put the email in the correct category. Not just some words, but the entire email, including links.

Of course there will be mistakes from time to time and the initial classification was pretty much inaccurate, but after sorting about 15,255 emails (I didn’t know I’ve had that many emails!) it has achieved an accuracy rating of 97.25%, which is impressive because I’ve set up 8 different possible categories for classification.

If you use Mozilla Thunderbird or MS Outlook and has been plagued by spam emails, or if you want a program to automatically categorise your emails, check out POPFile. You’re gonna love it.

Bored,
bcc

This entry was posted on Thursday, September 22nd, 2005 at 10:51 am and is filed under General. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.