基于改进谱聚类的正则表达式分组算法  

Regular Expression Grouping Algorithm Based on Improved Spectral Clustering

作　　者：陈曦[1] 陈庶樵[1] 刘大虎 CHEN Xi1, CHEN Shuqiao1 , LIU Dahu2(1National Digital Switching System Engineering ＆ Technology Research Center Zhengzhou 450002, China ; 2. Unit 68002, Lanzhou 730000, Chin)

机构地区：[1]国家数字交换系统工程技术研究中心 [2]68002部队

出　　处：《信息工程大学学报》2018年第1期95-99,共5页Journal of Information Engineering University

基　　金：国家973计划资助项目(2012CB315901;2013CB329014);国家863计划资助项目(2015AA016102;2013AA013505)

摘　　要：提出一种面向全局的正则表达式分组算法,即通过拉普拉斯矩阵将规则集合映射到具有明显聚类现象的空间中,将分组问题转化为传统的空间聚类问题,然后运用初始点优化的KMeans聚类方法实现快速分组。实验结果表明,在相同分组数的情况下,该算法的内存占用较GABG算法减少了10%左右,分组时间上缩短了2倍~3倍,实现了分组时间和分组效果的有效平衡。With the abundance of traffic and the development of detense technology, the existing reg- ular expression grouping algorithm is more and more difficuh to meet the growing storage demand. This paper puts torward a regular expression grouping algorithm with global orientation. Firstly, the rules set is mapped into the teature vector space with obvious clustering phenomenon by laplacian matrix. Then the improved K-Means clustering method is designed to realize fast grouping. Experi- mental results show that the algorithm can achieve the effective balance of grouping time and group- ing ettect. Compared with the current GABG algorithm, the algorithm can reduce 10% memory us- age for the same number of groups, and reduce the grouping time by two to three times.

关 键 词：正则表达式 深度报文检测 分组算法 谱聚类 拉普拉斯矩阵 K-MEANS聚类 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]