Serial and Parallel Bayesian Spam Filtering using Aho-Corasick and PFAC

Abstract

With the rapid growth of Internet, E-mail, with its convenient and efficient characteristics, has become an important means of communication in people‟s life. It reduces the cost of communication. It comes with Spam. Spam emails, also known as „junk e-mails‟, are unsolicited one‟s sent in bulk with hidden or forged identity of the sender, address, and header information. It is vital to pursue more effective spam filtering approaches to maintain normal operations of e-mail systems and to protect the interests of email users. In this paper we developed a Spam filter based on Bayesian filtering method using Aho-corasick and PFAC string matching algorithm. This filter developed an improved version of spam filter based on traditional Bayesian spam filtering to improve spam filtering efficiency, and to reduce chances of misjudgement of malignant spam. For further improvement of Spam filtering process we are transform the filter in to parallel spam filter on GPGPU's by using PFAC Algorithm.