Skip to Main Content
The paper described the basic principle of Naive Bayesian spam filter, and analyzed some major factors which influence Chinese spam filter mostly. On the basis of these, proposed a method to optimize the feature databases. Some experiments were designed to study the filter performance on correctness and efficiency. Some detailed analysis of the experiments also presented in this paper. The experimental results illustrate that the improved method can enhance the precision and efficiency of Bayesian classifier and lower the error rate of legitimate emails.