Research And Application Of Behavior Based Anti-Spam Technology

Email is one of the most successful applications of Internet, and it has become one important part of people's life, work and learning. While enjoying email's simpleness and convenience, people are suffering from its byproduct. That is spam overruns on internet. This not only brings users inconvenience but also is threatening internet's security. Therefore preventing spam is one internet problem that needs to be resolved quickly. Today avoiding spam still depends on anti-spam technologies.First, this paper studies and summarizes the existing anti-spam technologies. Although they have some effect on preventing spam, they appear some shortcomings as time goes, such as high incorrectly identifying ratio ,low disposal efficiency.Second, analyzing a mass of email samples, we find that the different essence between ham and spam must result in their different behaviors. This paper studies behavior filter technology deeply from spam behavior.1. Network topology based behavior filter technology is proposed. Email network topology graph is set up by user's mailbox and user's communication connection. Spam and ham are identified by using graph theory to analyze difference between spam and ham in topology graph.2. Sending behavior character based filter technology is put forward. Distinguishable sending behavior character from mail envelope and head is fully rediscovered and taken out, then befitting classifying arithmetic is selected to form classifying machine. Finally, email is identified by using sorting machine.Experiment result shows that behavior based filtering spam technology overcomes the shortcomings of the traditional technologies. And it has its own advantages such as high precision ratio, high efficiency and long life cycle.Finally, design and realize anti-spam gateway system. And behavior filter technology is also applied to it.