A Knowledge Acquisition Method of Judgment Rules for Spam E-mail by using Self Organizing Map and Automatically Defined Groups by Genetic Programming

Abstract

Recently, the Internet has been a basis of our cultural life. Web provides a virtual huge space of information, where an emotional experience, a political idea, a cultural custom, and the advice of manners of music, the business, the arts, photographs, and literatures, etc. are digitalized at a low price and are shared for our current culture. One of major other tools gives E-mail communication which propagates not only letters from person to person, but a means of advertisement. However, E-mail addresses on the web page are gained and the virus is appended to E-mails, and the attacks for acquiring information on user's personal computer (called BOT) have been spreaded. Their accumulated E-mail addresses collected in such a way will be valuable for advertising agents. But, their E-mails are called “Spamming“ and the Spamming becomes a one of the social issues. Spamming is the abuse of electronic messaging systems to send unsolicited bulk messages or to promote products or services, which are almost universally undesired. Spamming is economically viable because advertisers have no operating costs beyond the management of their mailing lists. The sender cannot be specified, because the sender of Spamming has only temporary E-mail address and the reply of them is not reached to the original sender. Therefore, undesired E-mails to us have been increased everyday, so that, it is not easy to read an important E-mail. In order to avoid Spam E-mails, we must build the filtering system which can judge whether the received E-mail is a Spam E-mail or not. The SpamAssassin(SpamAssassin) is the open source software and is a mail filter which attempts to identify a Spam E-mail using various pattern match methods including text analysis, Bayesian filtering, DNS block lists, and collaborative filtering databases. There are the predefined rules for each filtering method to detect a Spam E-mail. The agreement degree for each rule is scored and if the total score is larger than the threshold value, the given E-mail is judged as Spam. Although each score in the rule of SpamAssassin is low, there is a few cases in which total score becomes high. Moreover, even if the message is judged as Spam, the different rules for each person are 24