基于免疫的异常检测中实值自体集优化算法

Real-valued self set optimization algorithm in immunity-based anomaly detection

针对免疫异常检测一直被忽视的实值自体集多分区、样本重叠率高和噪声等现象,以及造成的检测器生成代价高和边界漏洞等问题,提出一种实值自体集优化算法。算法通过模糊聚类算法处理集合多分区问题,利用高斯理论对自体集中的噪声样本、高重叠率等问题进行处理。通过Iris数据集和网络数据验证,算法可以有效地解决以上问题,提高生成检测器的效率和系统检测率。

The real-valued self set in the immunity-based anomaly detection which is used to train detectors has some defects： multi-area,overlapping,noising sample,etc,which can cause some problems,such as the boundary holes of detector set,the high cost of generating detectors,etc.To solve the problems,this paper proposed a real-valued self-set optimization algorithm which used fuzzy clustering algorithm and Gaussian-distribution theory.The fuzzy clustering delt with multi-area and the Gaussian-distribution delt with the overlapping and noising.It tested algorithm by Iris data and real network data.Experimental results show that,the optimized self set can increase the efficiency of detector generation effectively,and improve the system＇s detection rate.