Content-Based Spam Filtering Model: Research And Realization

Email is becoming a convenient and rapid communicative tool. But more and more spams bring great problems, they would use up our resources, such as storage spaces in email servers,and they would affect the safety of our network. It is urgent to solve the"spam"problem. At present, the primary method of preventing spam is filtering.. Nowadays, anti-spam measurements often include black or white list technique, manual rules and keyword based content filtering. These measurements are becoming sick in filtering. Content-based filtering is becoming a hotspot. The typical method of content-based spam filtering is that of Bayesian Model based spam filtering.Basing on text categorization techniques and Bayesian theory, we established the email filtering model on Bayesian Classifier after analyzing the key techniques and methods of content-based filtering, and we provided the effective way and key algorithm in filtering and completed the judgment of spams .In order to reduce the damages of wrong e-mail classifying, we provided the improved methods of using the risk minimization Bayesian decision and self-adjustment of categorization system. We finally described a Spam Filtering model and realization of the prototype based on Bayesian model with related improvements.