Chinese Spam Filter System Based on Analysis Using Milter Interface

This paper presents a scheme of a real-time Chinese spam mail filtering system based on content analysis.The system works on Sendmail mail server under Linux.It utilizes Milter interface to get the real-time e-mail content,and then classifies and filters it combined with Chinese word segmentation and text categorization algorithms.It has high expansibility since it can embed many kinds of text categorization algorithms.Furthermore,these different text categorization algorithms are analyzed and compared by experiments.