文献类型：期刊文献

英文题名：A GPU-RSVM based intrusion detection classifier

作者：Zhang, Xueqin[1]; Zhao, Chen[1]; Wu, Jiyi[2]; Song, Chao[1]

机构：[1] School of Information Science and Engineering, East China University of Science and Technology, Shanghai, China; [2] Hangzhou Key Lab. of E-Bussiness and Information Security, Hangzhou Normal University, Zhejiang 310036, China

年份：2011

卷号：164 CCIS

起止页码：92

外文期刊名：Communications in Computer and Information Science

收录：EI(收录号：20114514505495)

语种：英文

外文关键词：Quadratic programming - Graphics processing unit - Intrusion detection

摘要：Recently support vector machines based intrusion detection methods are increasingly being researched because it can detect unknown attacks. But solving a support vector machine problem is a typical quadratic optimization problem, which is influenced by the number of training samples. Due to GPU's high performance in parallel computing, this paper proposes a Euclidean distance based reduction algorithm developed on GPU platform, which is called GPU-RSVM, to eliminate samples that have less effect on building SVM classifier. Experiment results show that the time of reduction process can decrease significantly. With optimal reduction ratio, the overall performance of the intrusion detection classifier based on the proposed GPU-RSVM algorithm is better than that based on LIBSVM algorithm. ? 2011 Springer-Verlag.