Research On Code Behavior-based Anti-virus Mechanism

People's life has been changing by the progress of Internet .But not only networks bring us great convenience but also bring us many kinds of security threat, in which Internet virus is the most important part. With the expanding of Internet ,the spread of Internet virus has become more and more quickly ,and damage to the people become larger and larger, so, how to defend Internet virus become a hotspot on the computer security research.Currently, most anti-virus technologies are feature-based. This framework is only effective to known virus, while it is useless for unknown ones. To solve the problem, this paper use data mining method to detect Internet virus, using the code behavior as eigenvector. We put forward an improved bayes algorithm for unknown virus detection.Based on the detecting method, this paper designed a behavior-based anti-virus system, and carry out the subsystem for unknown virus detection. The testing result of the system shows that the design and scheme are feasible.