基于数据挖掘的入侵检测系统设计  

Design of Data Mining Based Intrusion Detection System

作　　者：李守国[1] 李俊[1] 

机构地区：[1]南京航空航天大学计算机科学与工程系,江苏南京210016

出　　处：《计算机技术与发展》2006年第4期212-214,共3页Computer Technology and Development

基　　金：国防科工委国防基础科研项目(S0500B003)

摘　　要：现有的入侵检测系统大多都是采用手工编码构造的,检测模型的构造过程很大程度上依赖于系统构造者的知识和经验,这样构造出的模型往往存在很大的缺陷。针对传统入侵检测系统构造过程中存在的种种问题,将数据挖掘技术引入入侵检测系统,实现检测模型构造的自动化。介绍了一个运用数据挖掘技术构造入侵检测系统的框架,并考虑到实时检测过程中对检测模型效率的要求,提出了一个提高检测模型检测效率的层叠检测模块方法。应用数据挖掘算法得出的检测模型在检测效率、准确性、可扩展性和自适应性等方面都得到了很大的改进。Currently most intrusion detection systems are constructed by manual means, constructing process of these systems highly depend on system builders＇ knowledge and experience, so there are always some flaws in these models. Aiming at the problems exists in the process of IDS（intrusion detection system） constructing, apply data mining technology to IDS to construct the models automatically. In this paper, introduce a data mining based intrusion detection framework, and considering the real - time efficiency requirement of the intrusion detection model, propose ＂cascaded detection modules＂ approach to improve efficiency of IDS. The models constructed by applying data mining algorithms improve a lot in some aspects, such as efficiency, accuracy, extensibility and adaptability.

关 键 词：网络安全 入侵检测 数据挖掘 移动代理 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]