摘要

A new immune based model called AIBM for dynamic intrusion detection is proposed. AIBM uses a very small dynamic self set during the self tolerance for immature detectors, resulting in a higher efficiency in generating new mature detectors than traditional computer immune systems (CIS). Meanwhile, the self set can synchronize their variations with the real-network environment as time goes on, resulting in the dynamic evolution of self set, mature and memory detectors, offering more self-adaptation, and having a lower error rate than traditional CIS models.