一种面向IDS的检测器动态生成算法研究

ON DYNAMIC GENERATION ALGORITHM OF IDS-ORIENTED DETECTORS

针对目前基于免疫的入侵检测系统中报警可信性较低的问题,提出一种有效的检测器动态生成算法,并给出了其在入侵检测系统中的具体实现。该算法借鉴免疫系统的激活克隆和亲和力变异机制,能够依据实际情况不断调整当前检测器集合,并仅用较小的检测器集就能够快速检测到大规模非我空间中的异常变化。实验表明,算法用于入侵检测系统,提高了检测的准确率,降低了虚警率。同时,算法对各种异常检测向题具有一定的适用性。

An effective dynamic generation algorithm of detectors was presented aiming at improving the lower alarming creditability of current immunity-based intrusion detection systems,and its specific application on intrusion detection system was also proposed.The algorithm can regulate current detector sets continuously according to actual circumstances and quickly detects abnormal changes in a very large non-self space using a smaller detector sets by taking the mechanisms of activation-clone and affinity mutation in immune system as the reference.An experimental exam and its results prove that the algorithm improves the detection accuracy rate and reduces the false alarm rate when being used in intrusion detection system.At the same time,the algorithm can also hold for other anomaly detection problems.