Payload-based anomaly detection using KPCA

被引：0

作者：

Jia, Libin ^{[1
]}

Ma, Jun

Li, Lin ^{[1
]}

机构：

[1] Zhengzhou Inst Aeronaut Ind Management, Zhengzhou 450015, Peoples R China

来源：

PROCEEDINGS OF 2009 INTERNATIONAL CONFERENCE OF MANAGEMENT ENGINEERING AND INFORMATION TECHNOLOGY, VOLS 1 AND 2 | 2009年

关键词：

KPCA; Payload-based; Security;

D O I：

暂无

中图分类号：

TP39 [计算机的应用];

学科分类号：

081203 ; 0835 ;

摘要：

In this paper, we propose a new payload-based anomaly detector. N-gram technique is used to extracted feature from payload and feature reduction is fulfilled by Kernel Primacy Component Analysis (KPCA). In order to resist to mimicry attack, we don't only focus on the Primary Components but also make use of the residual components. In this manner, the normal payload can be described in a finer granularity. One-class SVM classifier is used to construct the normal model. Ensemble of one-class SVM classifiers is also employed to combine the different model in order to improve robustness of detector. Experimental results show that our detector achieve better performance than other payload-based anomaly detection systems.

引用

页码：566 / 569

页数：4