基于优化聚类算法的安全审计模型

论文针对网络未知入侵类型检测问题,提出了基于优化聚类算法的安全审计模型。由于攻击类型未知,无法设定聚类数目,这是影响聚类算法在日志分析中应用的主要问题。论文针对这种问题提出了优化的聚类算法,该算法由聚合阶段和优化阶段两部分组成,能自动实现初始聚类集合建立和组合优化,并通过简单的人工干预可准确地标识入侵行为。该算法处理未知入侵检测问题是可行,具有良好的可扩展性,大大降低了算法对控制参数的依赖性,提高了该方法的实用性。

Security Audit Model Based on Optimized Clustering Algorithm

In this paper author proposal a security audit model based on optimized clustering algorithm.Model can be described as two stages.First is stage of clustering,in this stage a common clustering algorithm has been used to get the wanted clusters.Second is stage of optimized,in this stage a strategy based on mean square difference has been used.This strategy is used to solve the uncertain number of cluster.At the end of this stage an intervener is involved to improve veracity of result.The result of examination shows this model has good veracity and useable.