摘要
构建计算机网络的入侵检测系统,对于保护网络中的信息免受各种攻击显得非常重要.为了克服支持向量机的缺点,提出了一种基于粒子群优化相关向量机(RVM)网络入侵检测方法.相关向量机是一种建立在支持向量机上的稀疏概率模型.与支持向量机相比,它不仅具有较高检测精度,还具有较好的实时性,粒子群优化算法用于确定相关向量机的核参数.最后结合试验将提出的方法同支持向量机算法、BP神经网络进行了比较,结果表明提出的相关向量机相比于支持向量机、BP神经网络有着更高的入侵精度.
It is significant to protect the information of network and avoid it under attack by constructing network intrusion detection system. In order to overcome the drawbacks of support vector machine,network intrusion detection based on relevance vector machine optimized by particle swarm optimization algorithm （PSO-RVM） is presented in the paper. Relevance vector machine is a sparse probability model based on support vector machine. Relevance vector machine has not only higher detection accuracy,but also better real-time than support vector machine. In the study,particle swarm optimization algorithm is used to determine nuclear parameter of relevance vector machine. Finally,the case data are used to testify and analyze the performance of the proposed model. The experimental results show that PSO-RVM has greater prediction accuracy than PSO-SVM,BP neural network.
出处
《微电子学与计算机》
CSCD
北大核心
2010年第5期181-184,共4页
Microelectronics & Computer
关键词
相关向量机
网络异常
入侵检测
粒子群
relevance vector machine
network anomaly
intrusion detection
particle swarm
