基于划分-移动的实值检测器分布优化研究
首发时间:2012-03-23
摘要:传统的免疫异常检测算法主要针对否定选择思想做出了一些改进,由于检测器生成具有随机性、匹配算法的一些固有特性,以及对于检测器泛化能力的要求等问题,导致了入侵检测系统中存在大量的漏洞。本文提出了一种实值检测器二次分布的启发式算法,它以检测的漏报数据为基础,对非自体空间的漏洞区域,通过划分-移动的方法进行实值检测器的二次分布。研究结果表明,该算法有效地避免漏洞的产生,降低了检测系统的漏报率。
For information in English, please click here
Research on Distribution Optimization of Real Value Detectors Based on Division and Movement
Abstract:To be aimed at the idea of negative selection, traditional anomaly detection algorithm makes some improvements. It exists many problems which are the randomness of detectors generation, some inherent characteristics of the matching algorithm and the generalization ability of detectors and so on. Therefore, it leads to a lot of vulnerabilities in intrusion detection system. This paper presents a heuristic algorithm based on the second distribution of real value detectors. For vulnerability areas of the non-self space, it distributes real value detectors through the method of division and movement which is based on dates of omission. The research shows that the algorithm effectively avoids the generation of vulnerabilities and reduces the omission rate of detection systems.
Keywords: network security vulnerability negative selection algorithm omission rate detector
论文图表:
引用
No.****
同行评议
共计0人参与
勘误表
基于划分-移动的实值检测器分布优化研究
评论
全部评论0/1000