一种基于数据挖掘的入侵检测方法研究与实现  

Study on the Realization of Detection of Intrusion Based on Data Mining

作　　者：刘茗[1] 

机构地区：[1]淮海工学院计算机工程学院,江苏连云港222005

出　　处：《淮海工学院学报（自然科学版）》2010年第3期16-20,共5页Journal of Huaihai Institute of Technology:Natural Sciences Edition

摘　　要：通过对入侵检测和数据挖掘技术的研究,介绍了传统入侵检测的类型和局限性,分析了数据挖掘技术在入侵检测系统中的应用。针对入侵检测领域的特点,指出利用数据挖掘技术可以克服这些局限性。通过对经典关联、聚类算法的改进和优化,从而解决了聚类算法固有的无法预知最佳聚类个数和分类过细的问题,因此降低了漏报和误报率,提高了入侵检测的效率。实验结果表明,该方法行之有效。This paper analyzed the application of data mining in intrusion detection system based on a research on IDS,data mining and the types and limitations of traditional detection methods.According to the characteristic of IDS,it was pointed out that these limitations could be overcome by the data mining technology.The intrusion detection technology improved and optimized the association and cluster rules,which resolved its own disadvantages,including its inability to presage the number of the best cluster and its fine classification.Such an improvement reduces omissions and misstatement,and improves the efficiency of IDS.Experimental results show that this algorithm is feasible.

关 键 词：入侵检测系统 数据挖掘技术 误用检测 关联规则 聚类算法 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]