基于Boosting与SVM的入侵检测技术  

Intrusion detection system based on Boosting and support vector machine

作　　者：牛永洁[1] 常浩[2] 

机构地区：[1]延安大学计算中心,延安716000 [2]太原大学计算机系,太原030009

出　　处：《信息技术》2010年第12期92-93,98,共3页Information Technology

摘　　要：为克服目前很多入侵检测方法存在成功率低以及误警率高的缺点,将Boosting与SVM算法结合,使用小训练样本对SVM进行训练,得到分类器,然后使用Boosting方法进一步提高SVM的泛化能力。在Matlab 2009版本下,采用KDD99入侵检测数据进行了仿真实验。仿真实践证明,这种技术可提高分类精度和准度,提高了入侵检验的成功率。For improving attack detection rate and lowering false positive rate of many current intrusion detection methods,Boosting algorithm is combined with support vector machine,this paper makes use of support vector machine to train and get classifier by small training samples,then uses the Boosting algorithm to further improve the generalization ability of support vector machine.In Matlab 2009,the use of KDD99 intrusion detection data were simulated.The simulation proved that this technology can improve the classification precision and accuracy,and improve the success rate of intrusion detection.

关 键 词：入侵检测 BOOSTING 支持向量机 仿真 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]