Real-time statistical rules for spam detection

Spam detections fall into two categories: rule-based and statistical-based. The former refers to the detection which is performed by looking for spam-liked patterns in an email. Since the rules can be shared, they have been popularized quickly. The rules, however, are built manually it is hard to keep them up with the variation of spam. The statistical-based method, on the other hand, is possible to make the detector retrained quickly, but knowledge obtained from this method is unable to be shared among the servers. We, therefore, proposed a statistical rule-based method for spam detection. A widely used rule set Chinese_rules.cf, for SpamAssassin to catch spam written in Chinese is generated by this method. It can be updated automatically and can also be shared among servers. A generating process of the Chinese_rules.cf is described. Factors that control the rule’s performance are discussed.