基于灰狼算法优化DBN的医院网络异常流量识别  

Recognition of Abnormal Traffic in Hospital Network Based on Gray Wolf Algorithm to Optimize DBN

作　　者：黄波 杨正 王超[2] HUANG Bo;YANG Zheng;WANG Chao(Guangzhou Eighth People’s Hospital, Guangzhou 510060, China;College of Optoelectronics and Information Technology, Fujian Normal University, Fuzhou 350007, China)

机构地区：[1]广州市第八人民医院,广东广州510060 [2]福建师范大学光电与信息科技学院,福建福州350007

出　　处：《微型电脑应用》2022年第1期34-36,44,共4页Microcomputer Applications

基　　金：广东省科技厅支持项目(17G9812)。

摘　　要：为了提高医院网络异常流量识别的精度,提出一种基于灰狼算法优化DBN的医院网络异常流量识别方法。针对DBN模型性能受权值和偏置参数的影响,运用灰狼算法对DBN模型的权值和偏置进行优化选择,将医院网络流量特征数据作为DBN模型的输入向量,网络异常流量的类型作为DBN模型的输出向量,建立GWO-DBN的医院网络异常流量识别模型。研究结果表明,GWO-DBN进行医院网络异常流量识别具有更高的准确率、检测率和更低的误报率。In order to improve the accuracy of identifying abnormal traffic in hospital network,a method for identifying abnormal traffic in hospital network based on gray wolf algorithm optimization DBN is proposed.As performance of the DBN model is affected by the weights and bias parameters,the gray wolf algorithm is used to optimize the weights and biases of the DBN model.The hospital network traffic characteristic data are used as the input of the DBN model,and the type of abnormal network traffic is regarded as the output of the model.The output vector is used to establish the GWO-DBN hospital network abnormal traffic recognition model.The research results show that the GWO-DBN has higher accuracy,detection rate and lower false alarm rate for identifying abnormal traffic in hospital network.

关 键 词：深度置信网络 灰狼优化算法 网络流量 受限玻尔兹曼机 

分 类 号：R197.324[医药卫生—卫生事业管理] TP393[医药卫生—公共卫生与预防医学]