A Dynamic Immunity-Based Model for Computer Virus Detection

Internet provides a fertile medium for new breeds of computer viruses. Many people who have access to a wealth of information via Internet are attacked by more computer viruses than they can effectively process. We present a dynamic computer virus detection model that can detect known viruses and previously unknown viruses to prevent information systems from damage. This model is inspired by biological immune systems that protect the body against damage from pathogens. The architecture of this model, the formal definitions of self, nonself, antigen, antibody, and vaccine gene library are introduced. Furthermore, the evolution of self and nonself, the generation of the antibody, the evolution of the virus vaccine gene, and the detection of the antigen are depicted. Experiment results show that this model has better capacity of self-adaptability and self-learning in detecting unknown viruses than traditional models.