改进聚类算法在DRDoS攻击检测中的应用研究  

Research and Application of Improved Clustering Algorithm in DRDoS Attack Detection

作　　者：温海波[1] WEN Haibo(Department of Urban Rail Transit and Information Engineering,Anhui Vocational and Technical College of Communications,Anhui Hefei,230051,China)

机构地区：[1]安徽交通职业技术学院城市轨道交通与信息工程系,安徽合肥230051

出　　处：《安徽建筑大学学报》2017年第1期70-75,共6页Journal of Anhui Jianzhu University

基　　金：安徽省教学研究项目(2016mooc126);安徽省自然科学基金重点项目(KJ2016A159)

摘　　要：针对DBSCAN及K-means聚类算法固有缺陷,提出基于信息熵的贪婪迭代增长式聚类算法(GIG-CA)。该算法将熵值理论引入聚类初始点的确立,以类内对象间欧几里德距离均值作为聚类的区域密度阈值,用贪婪法迭代增长生成聚类。将该算法应用于DRDoS攻击检测中,通过实验验证其正确性、鲁棒性及实用性。In this paper, a greedy iterativ e growth clustering algorithm (GIGCA) based on information ent ropy is proposed for DBSCAN and K-means clustering algorithm. The entropy theory is introduced into the initial point of clustering. The Euclidean distance between objects in the class is taken as the regional density threshold of clustering, and the clustering is generated by greedy method. The algorithm is applied to DRDoS attack detection, and it s correctness, robustness and practicability are verified by experiment s.

关 键 词：聚类算法 信息熵 区域密度 贪婪迭代增长 入侵检测 DRDOS 

分 类 号：TP393.8[自动化与计算机技术—计算机应用技术]