Using Artificial Neural Network to Filter Spam for Chinese Mail

As the Internet developed, the problem of spam has become increasingly serious. Not only caused great distress to individuals, but also have a great business costs. With improvements in computing speed, neural network is becoming a very good tool for text classification. The purpose of this study is to conduct few experiments by using neural network approach for Chinese mails’ content. The result shows that neural network approach is effective for Chinese mails’ spam-identification and the adjustments of some parameters (the number of keywords, the number of nodes, and the number of categories) also increase the accurate rate, while reducing false positives.