基于改进决策树的5G网络环境中非法行为自动化检测系统  

An automated detection system for illegal behavior in 5G network environment based on improved decision tree

作　　者：雷琳琳 喻朝新 李丹 LEI Linlin;YU Chaoxin;LI Dan(China Mobile GBA(Greater Bay Area)Innovation Institute,Guangzhou 510623,China)

机构地区：[1]中国移动大湾区(广东)创新研究院,广东广州510623

出　　处：《电子设计工程》2025年第7期155-158,163,共5页Electronic Design Engineering

摘　　要：在5G网络环境中,非法行为的发生对5G网络安全构成了严重威胁,当前的决策树算法受限于流量限制和预警数据冗余问题。设计基于改进决策树的5G网络环境中非法行为自动化检测系统。基于B/S设计的系统三层框架,设计四个功能单元,采集单元采集网络流量数据,特征提取单元分析网络流量数据并提取IMF分量的能量特征,改进决策树检测单元利用ID3算法建树,实现自动化检测,预警单元根据检测到的非法行为生成预警信息并及时通知管理员,实现预警。结果表明,所设计系统在应对四种不同非法行为时,杰卡德系数均达到0.9以上。In the 5G network environment,the occurrence of illegal behaviors poses a serious threat to 5G network security,and the current decision tree algorithm is limited by traffic restriction and early warning data redundancy.An automatic detection system for illegal behavior in 5G network environment based on improved decision tree is designed.A three-layer system framework based on B/S design.Four functional units are designed:the collection unit collects network traffic data;the feature extraction unit analyzes network traffic data and extracts the energy characteristics of IMF components;the decision tree detection unit is improved to build trees using ID3 algorithm to realize automatic detection;the early warning unit generates early warning information based on detected illegal behaviors and timely informs the administrator to realize early warning.The results show that the Jacard coefficient of the designed system is above 0.9 when dealing with four different illegal acts.

关 键 词：改进决策树 5G网络环境 非法行为 自动化检测系统 

分 类 号：TN98[电子电信—信息与通信工程]