Mimicry Detection Method Based on Immune Danger Theory

Abstract

Considering that the traditional detector has lower detection rate, resource utilization and higher false alarm rate, this paper proposes a mimicry algorithm and a mimicry detector based on Immune Danger Theory (IDTMD). The detector adjusts the types and number of detection organizations and detection cells dynamically, according to the types and number of the detected signal and the detected dangerous signal, as to realize the dynamic, diversity and randomness of mimicry detection. The KDD-CUP99 dataset was used to test the comprehensive detection performance of the IDTMD detector, the detector based on Negative Selection Algorithm (NSA), the detector based on Dendritic Cell Algorithm (DAC). The results show that the detector improves the detection rate and the resource utilization, decreases the false alarm rate, and when testing amount reaches a certain scale, its response time is significantly less than other detectors’.