1. Introduction
Electronic mail is an important communication method for most computer users. The amount of email is increasing all the time, and in particular, the amount of unsolicited commercial email-spam-has increased dramatically over the last few years. Users typically distinguish spam from non-spam, and organize their email into a hierarchy of mail folders. Email clients typically allow users to manually move email to folders through some user interface, and also provide a facility for a user to define custom email processing rules. These rules are limited in power, and manually organizing email soon becomes a very tedious job. To assist a user to cope with this work load, several approaches have been followed, ranging from spam detection software to email assistants to aid the mail organization process. Organisation of mail then becomes a classification task, where non-spam email must be assigned to the ‘correct’ folder.