关于通信无线网络异常账号识别仿真研究  

About Communication Wireless Network Anomalies Account Recognition Simulation Research

作　　者：李刚[1] 李锐君[2] 

机构地区：[1]平顶山学院现代教育技术中心,河南平顶山467000 [2]郑州大学西亚斯国际学院,河南郑州451150

出　　处：《计算机仿真》2016年第8期268-271,321,共5页Computer Simulation

摘　　要：对通信无线网络异常账号进行准确识别,可提升公共通信无线网络安全性。异常账号识别时,需要通过建立异常账号特征选择的数学模型,获得无线网络异常账号最优特征和分类器参数,而传统的分层概率图算法利用统计假设方法对公共通信无线网络异常账号检测,不能建立精确的账号特征选择模型,降低了异常账号检测的精度。提出一种关于特征和分类器联合优化的公共通信无线网络异常账号检测方法。上述方法先建立公共通信无线网络异常账号特征选择的数学模型,采用遗传算法获取公共通信无线网络异常账号特征子集,将网络状态特征和SVM分类器参数作为遗传算法的个体,以网络异常账号检测正确率作为个体适应度函数,通过选择、交叉和变异等遗传操作获得无线网络异常账号最优特征和分类器参数,并有效地完成了对公共通信无线网络异常账号检测。仿真结果证明,改进算法能提高无线网络异常账号检测的效率、稳定性、精度。In order to increase detection precision,a detection method of abnormal accounts is proposed based on combined optimization of feature and classifier in wireless communication network. Firstly,a mathematical selection model of abnormal account feature is built. Then the genetic algorithm is used to obtain character subset of abnormal accounts. The network state feature and SVM classifier are used as unit of genetic algorithm,and the detection accuracy is taken as the individual fitness function. Finally,the optimal feature and classifier parameter are obtained through genetic operation such as choosing,crossing and varying,and the detection is effectively completed. The simulation results show that the modified algorithm can improve efficiency,stability and precision of abnormal account detection in wireless network.

关 键 词：异常账号检测 公共通信 无线网络 

分 类 号：TD393.08[矿业工程—矿井建设]