基于机器学习的VoIP流量在线识别系统设计与实现研究  

Machine learning online recognition system design and implementation based on VoIP flow

作　　者：安强强 张峰[1] 李赵兴[1] 张雅琼 AN Qiang-qiang;ZHANG Feng;LI Zhao-xing;ZHANG Ya-qiong(Yulin University,Yulin 719000,China)

机构地区：[1]榆林学院,陕西榆林719000

出　　处：《电子设计工程》2018年第16期103-106,共4页Electronic Design Engineering

基　　金：陕西省教育厅2017年专项科学研究计划(17JK0900);榆林学院2016年高层次人才科研启动基金项目(16GK25)

摘　　要：本次以基于机器学习的VoIP流量在线识别系统设计与实现为研究目标。通过借助于数据挖掘工具WEKA集成机器学习算法,结合流量特征,筛选构建专门的模型,再设计一边抓包一边检测的机制,最终达到VoIP的快速识别。具体试验研究结果显示,该系统的在线识别率较高,可以达到的精度为91.4%。得出以机器学习为基础设计相应的VoIP流量在线系统能够实现对流量的在线高效识别,效果显著,可以推广应用。The design and implementation of VoIP traffic online identification system based on machine learning is the goal of this study.By means of the data mining tool WEKA integrated machine learning algorithm,combined with the traffic characteristics,select and construct the special model,then design mechanism of side side capture detection,and ultimately achieve rapid identification of VoIP.The experimental results show that the online recognition rate of the system is high,and the precision can reach 91.4%.It is concluded that the VoIP online traffic system based on machine learning can realize on-line and efficient identification of traffic,and the effect is remarkable,and it can be popularized and applied.

关 键 词：机器学习 VOIP 流量在线识别系统 设计实现 

分 类 号：TN99[电子电信—信号与信息处理]