An Improved Bayesian with Application to Anti-Spam Email
ZHANChuan, LUXian-liang, ZHOUXu, HOUMeng-shu · 中国电子科技：英文版 · 2005
Along with the wide application of e-mail nowadays, many spam e-mails flood into people's email-boxes and cause catastrophes to their study and life. In anti-spam e-mails campaign, we depend on not only legal measures but also technological approaches. The Bayesian classifier provides a simple and effective approach to discriminate classification. This paper presents a new improved Bayesian-based anti-spam e-mail filter. We adopt a way of attribute selection based on word entropy, use vector weights which are represented by word frequency, and deduce its corresponding formula. It is proved that our filter improves total performances apparently in our experiment.