Design and Implementation of an Optimization System of Span Filter Rule Based On Neural Network

One of the drawbacks of content-based filtering technology is that the system cannot adapt the filter to identify emerging spam characteristics. This paper describes the design and implementation of a spam filtering rules optimization system by introducing BP neural network. It can automatically extract features from incoming emails and "learn" so as to modify the filtering rules to accommodate new changes. We compare the performance of our system with spam assassin. Our experiment results show that the accuracy rate reaches 98.65%.