Research On Intrusion Detection Algorithm Based On Artificial Immune System

Network security is an important part of national security. As an intrusion detection system, artificial immune system evolved from bionics has drawn much attention. Artificial immune system is an effective active defense means, which provides a new way to research on intelligent intrusion detection.The main work of this paper includes the following two parts:1?In this paper a threshold method is proposed to deal with the detection of "black hole" problem,which is cased by r-bit continuous matching rule. And according to the threshold is an ordered sequence and meet the conditions of the binary search principle. The best matching threshold value of the detector is obtained by using the thought of the binary search principle. Then the fuzzy matching method is used to generate the mature detectors and detect the anomaly.2?Based on real valued positive selection algorithm using hyper sphere to describe the detector, in the face of the self set is large, computational overhead is unable to meet the current detection problem. In this paper k-means clustering is used to classify self-set collection and select a representative class center as a detector. This method can not only ensures the performance of detectors, but also greatly reduces the number of detectors. In addition according to the characteristics of network traffic,the detector test sequence given the dynamic priority so that detection system has maintained a higher efficiency and anomaly detectors are added to the detector set. The detection system also can detect the known intrusion quickly from the elements of self-set collection.