A spam analysis system

Abstract

With the popularization of the Internet, e-mail services become one of the tools of mass communication. In the same time, spam also annoys with people. Although there are many methods of analysis for spam, most of them only serve local users, and thus cannot provide more information about spam. Besides, single method can’t detect all types of spam. This paper provides an integrated e-mail security service system which combines the distributed log analysis, Bayesian classifier, url check, etc. Mechanism integrated from these methods and parameters adjusted via experiments make this system be able to detect more types of spam. By using forward and auto-reply functions, this system provides an e-mail analyzing service to public users, thus allowing collecting and processing various e-mail messages to improve the precision of spam detection.