马尔科夫聚类改进敏感信息自动监测  

Automatic Monitoring of Sensitive Information Based on Markov Clustering

作　　者：杜军龙[1] 周剑涛 王磊 DU Jun-long;ZHOU Jian-tao;WANG Lei(Jiangxi Information Center,Jiangxi Nanchang330036,China)

机构地区：[1]江西省信息中心,江西南昌330036

出　　处：《机械设计与制造》2020年第5期134-137,共4页Machinery Design & Manufacture

基　　金：广东省教育厅特色创新项目(自然科学)(2016KTSCX174)。

摘　　要：异常节点监影响通信网络路由性能,其敏感信息有利于发现网络攻击行为。为此,针对通信网络异常节点及其敏感信息监测,提出基于马尔可夫聚类改进的通信网络异常节点敏感信息监测方法。算法基于网络采集并预处理的流数据建立能够表征网络状态的邻接矩阵;然后在分析敏感词距离及敏感信息敏感度基础上,对邻接矩阵进行马尔可夫聚类处理,根据核心聚类节点在聚类前后的结构差异,实现通信网络中的敏感信息的自动监测。实验结果验证了算法在监测识别敏感信息方面的有效性和准确率。Abnormal node monitoring affects the routing performance of the communication network,and its sensitive information is conducive to discovering network attack behavior. So,The problem of automatic monitoring of sensitive information in communication networks is studied,and a new method of automatic monitoring of sensitive information in networks is proposed. In this method,an adjacency matrix capable of characterizing the network state is established based on the stream data collected and preprocessed in the network,and then,based on the analysis of sensitive word distance and sensitivity of sensitive information,Markov clustering is performed on the adjacency matrix. Automatic monitoring of sensitive information in communication networks is achieved according to the structural differences of the core cluster nodes before and after clustering. The experimental results verify the effectiveness and accuracy of the algorithm in monitoring and identifying sensitive information.

关 键 词：通信网络异常 敏感信息监测 马尔科夫 网络流聚类 

分 类 号：TH16[机械工程—机械制造及自动化] TP393[自动化与计算机技术—计算机应用技术]