基于攻击特征动态增量聚类的复合连续网络攻击安全检测方法

王远翔

基于攻击特征动态增量聚类的复合连续网络攻击安全检测方法

引用本文：王远翔.基于攻击特征动态增量聚类的复合连续网络攻击安全检测方法[J].计算技术与自动化,2025,(3):164-171

摘要点击次数: 291

全文下载次数: 198

作者	单位
王远翔	（湖南省金盾信息安全等级保护评估中心有限公司,湖南长沙 410000）

中文摘要:复合连续网络中，信息共享残差值过大是网络体系无法准确检测攻击性行为的主要原因。为提升复合连续网络运行安全性，实现对攻击性行为的精准检测，设计了基于攻击特征动态增量聚类的复合连续网络攻击安全检测方法。生成攻击性数据的分布序列，通过转换数据样本格式的方式，完成对复合连续网络中攻击特征的提取。在此基础上，针对复合连续网络进行动态层次实施划分，并在聚类结构中，计算攻击数据的动态增量聚类相异度，实现面向攻击特征的复合连续网络攻击数据动态增量聚类处理。设置必要的安全性攻击检测节点，联合节点内采集到的攻击信息，执行网络攻击数据的安全融合，完成复合连续网络攻击安全检测。测试结果表明，该方法可将信息共享残差控制在0-1.32之内，且能够准确高效地检测攻击性行为，在提升网络体系运行安全性方面具有突出应用价值。

中文关键词:攻击特征动态增量聚类复合连续网络攻击安全检测分布序列格式转化动态层次相异度

A Security Detection Method for Composite Continuous Network Attacks Based on Dynamic Incremental Clustering of Attack Features

Abstract:In composite continuous networks, the large residual value of information sharing is the main reason why the network system cannot accurately detect aggressive behavior. To improve the operational security of composite continuous networks and achieve precise detection of aggressive behavior, a dynamic incremental clustering based attack feature based composite continuous network attack security detection method is designed. Generate a distribution sequence of aggressive data, and complete the extraction of attack features in composite continuous networks by transforming the data sample format. On this basis, dynamic hierarchical partitioning is implemented for composite continuous networks, and the dynamic incremental clustering dissimilarity of attack data is calculated in the clustering structure to achieve dynamic incremental clustering processing of attack data for composite continuous networks oriented towards attack features. Set up necessary security attack detection nodes, combine the attack information collected within the nodes, perform security fusion of network attack data, and complete composite continuous network attack security detection. The test results show that this method can control the information sharing residual within 0-1.32, and can detect accurately and efficiently aggressive behavior, which has outstanding application value in improving the operation security of network system.

keywords:attack features dynamic incremental clustering composite continuous network attacks safety testing distribution sequence format conversion dynamic hierarchy degree of dissimilarity

查看全文 查看/发表评论 下载pdf阅读器