基于混沌PSO算法优化RBF网络入侵检测模型  

Based on Chaos PSO algorithm optimize RBF network intrusion detection

作　　者：王亚[1,2] 熊焰[1] 龚旭东[1] 陆琦玮[1] 

机构地区：[1]中国科技大学计算机科学与技术学院,合肥230027 [2]阜阳师范学院计算机与信息学院,安徽阜阳236037

出　　处：《计算机工程与应用》2013年第10期84-87,共4页Computer Engineering and Applications

基　　金：国家自然科学基金(No.61170233;No.61232018;No.61272472;No.61272317);全国统计科研计划资助项目(No.2012LY009);安徽省优秀青年人才基金项目(No.2011SQRL173)

摘　　要：针对网络安全中异常入侵检测,给出了一种构建最优神经网络入侵模型的方法。采用混沌扰动改进粒子群优化算法,优化径向基函数RBF神经网络入侵模型。把网络特征子集和RBF神经网络参数编码成一个粒子,通过粒子间的信息交流与协作快速找到全局最优粒子极值。在KDDCup99数据集进行仿真实验,实验数据表明,建立了一种检测率高、速度快的网络入侵检测模型。For anomaly intrusion detection in network security, this paper proposes a method of establishing the optimal neural network intrusion model. It improves particle swarm optimization algorithm by chaos perturbation. And it optimizes Radial Basis Function（RBF）neural network intrusion model. The subset features of network and RBF neural network parameters are considered as a particle. It uses the inter particle exchange of information and collaboration to find the global optimal particle extremum quickly. The simulation experiment is carried out on KDD Cup99 datasets. The simulation results show that it is a high detection ratio and fast speed network intrusion detection model.

关 键 词：入侵检测模型 特征选择 粒子群优化算法 神经网络 混沌扰动 数据集 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]