With the accelerated expansion of the Internet, people’s everyday life has become more and more inseparable from the various services supported by the Internet. E-mail is one of the most popular services on the Internet. As one of the earliest Internet services, e-mail has changed the way people exchange message so that the exchanges between the people become an unprecedented swiftness. But the attendant problem----spam, had brought businesses and individuals great losses and endless trouble. Spam develops rapidly in China. China has become the world third largest source of spam. At present, the study of the filter of the Chinese spam is in full swing. This paper gives some studies of the Bayesian filter of the Chinese spam.Bayesian spam filter is currently the most widely used, and it works better than every other spam filters. In this paper, aiming at the disadvantages of the traditional Bayesian filter, the author improves Bayesian model by importing attribute weight, threshold and timestamp. The author completed the design of a multi-layer integrated spare filtering system. According to the test result, the precision and recall of this filtering system is higher than the traditional filtering system based on the simple Bayesian model. |