最小差异度聚类在异常入侵检测中的应用  

Minimal diversity clustering applied in abnormity intrusion detection

作　　者：张应辉[1] 饶云波[1] 

机构地区：[1]成都东软信息技术学院计算机系,成都611844

出　　处：《计算机应用研究》2007年第12期193-195,225,共4页Application Research of Computers

摘　　要：重点研究了异常入侵检测系统模型。针对现有模型中存在的对训练数据要求高、误报率高等问题,提出了一种基于最小差异度聚类的入侵检测方法。该方法将区间标量、序数变量、二元变量标称变量类型的属性映射到区间[0,1]上,计算每个数据对象之间以及与各个簇的差异度,很好地解决了异常入侵。在检测已知入侵方面,模型也有不俗表现。This paper focused on the anomaly instrusion detection model. As the exisiting model causes the problems in misused detection and anomaly detection, proposed an clustering intrusion detection method based on minimum diversity and designed the intrusion detection module after the introduction and analysis of the clustering method. This method mapped [ O, 1 ] by scalar quantity, ordinal variable, duality variable. The intrusion detection was well solved by calculating the minimum diversity between in data object and individually. The model also could do well in detecting known intrusions.

关 键 词：网络安全 入侵检测 聚类分析 差异度 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]