基于主成分分析的高速网络IDS性能研究  

Research of high-speed network IDS based on principal component analysis

作　　者：马武[1] 陈波[1] 潘成胜[1] 

机构地区：[1]大连大学信息工程学院,辽宁大连116622

出　　处：《计算机工程与应用》2010年第10期82-84,87,共4页Computer Engineering and Applications

基　　金：辽宁省科技厅项目(No.2008403002)

摘　　要：随着高速网络的快速发展,如何在高速网络中快速有效地捕捉到异常的攻击特征,成为研究IDS所面临的首要问题。利用主成分分析技术的不同主成分互不相关和主成分是原始特征的线性组合的特性,有效地将高维特征向量映射到低维的空间中,既保持了原始数据的特征,又减少了高速网络环境下系统的丢包率,通过对KDDCup99数据集进行实验,并运用BP神经网络分类器进行了验证,证明该方法是正确有效的。同时提出了数据管理功能模块,不但使算法与实际应用结合的更加紧密,而且也改善了入侵检测系统的整体性能。With the rapid development of high-speed network,how to capture the unusual characteristics of the attack quickly and effectively in high-speed network has become an important problem of the IDS.Using the character of principal component analysis of the various principal component is not relevant and the principal component is linear combination of all the original features,the high-dimensional feature vectors are effectively mapped to the low-dimensional space.It has not only maintained the characteristics of the original data,but also reduced the discarded package rates under the high-speed network system.Through testing the experiment of KDD Cup99 data and using of BP neural network classifiers are proved that the method is correct and effective.Meanwhile,the data management capabilities module is designed,it makes the algorithm and the practical application more closely and improves the intrusion detection system＇s overall performance.

关 键 词：入侵检测 信息安全 主成分分析 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]