Learning how to tell ham from spam

Abstract

Unsolicited Commercial Email (UCE), commonly known as spam, has evolved from a mere nuisance to a multi-billion dollar problem. The near zero cost of acquiring huge lists of email addresses and flooding them with advertising messages has affected individual users, small companies, and large institutions alike. Spam is forcing users to wade through their mailbox to find the relatively few emails specifically addressed to them, colloquially referred to as "ham." Spam also wastes bandwidth, intermediate storage space, CPU time of Internet Service Providers (ISPs), and is often insulting and unsuitable (e.g. having pornographic content), especially to minors. Indicative of spam's extent is the first Conference on Email and Anti-Spam (CEAS) that recently took place in Mountain View, CA, following several industrial conferences on the topic. The conference attracted over two hundred academics and practitioners, most of them actively working on the problem from different perspectives, such as machine learning, security, and law.