摘要
为解决常用光纤网络异常节点数据挖掘方法耗时长、精度低的问题,提出一种可应用于大规模光纤网络的异常节点数据深度挖掘方法。预处理光纤网络节点数据,提取信息熵特征,并对数据进行降维操作,引入随机森林算法,通过自助采样形成多个随机样本空间,通过投票机制合并处理并输出树群中各棵子树光纤网络异常节点数据深度挖掘结果,实现光纤网络异常节点挖掘。实验结果表明,所提方法的精确度高达99.8%,耗时仅为9.2 min,漏检率为0.12%,因此,该方法可以获取高效率、高精度的光纤网络异常节点数据深度挖掘结果。
In order to solve the problem of time-consuming and low precision of common data mining methods for abnormal nodes in optical fiber networks,a deep mining method for abnormal nodes in large-scale optical fiber networks is proposed.Preprocess the data of optical fiber network nodes,extract the information entropy feature,and reduce the dimension of the data.Introduce the random forest algorithm,form multiple random sample spaces through self-help sampling,merge and output the data mining results of abnormal optical fiber network nodes of each sub tree in the tree group through the voting mechanism,so as to realize the mining of abnormal optical fiber network nodes.The experimental results show that the accuracy of the proposed method is as high as 99.8%,the time-consuming is only 9.2 min,and the missed detection rate is 0.12%.Therefore,this method can obtain high-efficiency and highprecision data mining results of abnormal nodes in optical fiber network.
作者
耿德志
宫海晓
GENG Dezhi;GONG Haixiao(Department of Information Technology and Engineering,Jinzhong College,Jinzhong 030619,China;School of Data Science&Software Engineering,Wuzhou University,Wuzhou 543002,China)
出处
《激光杂志》
CAS
北大核心
2023年第4期124-128,共5页
Laser Journal
基金
广西高校中青年教师基础能力提升项目(No.2020KY17019)。
