Spam Mail Filtering Based on Network Processor

As the rapid development of the Internet, the occurrence of more and more spam mails becomes harmful to users. Content-based spam filtering technologies become the mainstream anti-spam mail methods so far. Support vector machine (SVM), Bayes, windows and KNN are excellent ones of these technologies and they have advantages and disadvantages respectively. The common shortage of content-based methods is that they can´t filter spam mails as far as white-list-based, black-list-based or rule-based methods. This paper proposes a spam mail filtering system based on SVM and Bayes which is implemented on network processor (NP). To content-based methods, this system preserves the filtering accuracy and takes advantage of the parallel processing abilities of NP to improve the filtering speed.