Application and Evaluation of Bayesian Filter for Chinese Spam

Abstract

Recently, a statistical filtering based on Bayes theory, so-called Bayesian filtering gain attention when it was described in the paper “A Plan for Spam” by Paul Graham, and has become a popular mechanism to distinguish spam email from legitimate email. Many modern mail programs make use of Bayesian spam filtering techniques. The implementation of the Bayesian filtering corresponding to the email written in English and Japanese has already been developed. On the other hand, few work is conducted on the implementation of the Bayesian spam corresponding to Chinese email. In this paper, firstly, we adopted a statistical filtering called as bsfilter and modified it to filter out Chinese email. When we targeted Chinese emails for experiment, we analyzed the relation between the parameter and the spam judgement accuracy of the filtering, and also considered the optimal parameter values.