基于并行深度卷积神经网络的移动通信网突发流量异常检测方法  

Detection Method of Burst Traffic Abnormality in Mobile Communication Network Based on Parallel Deepconvolutional Neural Network

作　　者：王璇 赵克勤 WANG Xuan;ZHAO Keqin(College of information engineering,Zhongyuan institute of science and technology,Xuchang Henan 461000 China)

机构地区：[1]中原科技学院,信息工程学院,河南许昌461000

出　　处：《长江信息通信》2025年第2期43-45,共3页Changjiang Information & Communications

摘　　要：针对当前移动通信网突发流量异常检测存在查全率和交并比较低的问题,无法达到预期的检测效果,提出基于并行深度卷积神经网络的移动通信网突发流量异常检测方法。采用网络爬虫技术爬取通信网突发流量数据,并对其聚合、标识预处理,通过对数据主成分分析降低原始数据维度,采用并行深度卷积神经网络技术对突发流量数据异常特征提取和融合,识别检测到突发异常流量,实现基于并行深度卷积神经网络的移动通信网突发流量异常检测。经实验证明,设计方法查全率在95%以上,可以实现对移动通信网突发流量异常精准检测。In view of the problems of recall and low traffic detection,and the expected detection effect cannot be achieved,the detection method of sudden traffic detection in mobile communication network based on parallel deep convolutional neural network is proposed.Using network crawler technology to climb communication network sudden flow data,and the aggregation,identification pretreatment,reduce the original data dimension,through the parallel depth convolutional neural network technology of sudden flow data abnormal feature extraction and fusion,identify the detected sudden abnormal flow,realize the mobile communication network based on parallel depth convolutional neural network sudden traffic abnormal detection.The experiment proved that the recall rate of the design method is above 95%,which can realize the accurate detection of abnormal sudden traffic of the mobile communication network.

关 键 词：并行深度卷积神经网络 移动通信网 突发 流量 异常检测 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]