Adaptive spam mail filtering system

There are many techniques for filtering spam mails. Centroid-based classification method is one of the content-analysis techniques. It checks through every incoming mail and filters the spam mails with the frequently used terms, as vector-space method classifies the spam mail without considering the sequence of word. Those is method has the accuracy of filtering the spam mail. In this paper, we propose Adaptive Spam mail Filtering System which complements the mentioned weak point. The proposal system is combined of the 3 step filtering methods, such as pattern matching, two keyword matching, and centroid-based classification.