基于并行DAGSVM的网络流量分类方法  

Network traffic classification method based on parallel DAGSVM

作　　者：梁文国 王勇[2] 俸皓[2] LIANG Wen-guo;WANG Yong;FENG Hao(School of Information and Communication,Guilin University of Electronic Technology,Guilin 541004,China;Guangxi Colleges and Universities Key Laboratory of Cloud Computing and Complex Systems, Guilin University of Electronic Technology,Guilin 541004,China)

机构地区：[1]桂林电子科技大学信息与通信学院,广西桂林541004 [2]桂林电子科技大学广西云计算与复杂系统高校重点实验室,广西桂林541004

出　　处：《计算机工程与设计》2018年第2期316-319,335,共5页Computer Engineering and Design

基　　金：国家自然科学基金项目(61172053;61662018;61661015);广西自然科学基金项目(2016GXNSFAA380153);广西云计算与复杂系统高校重点实验室基金项目(2015209);广西研究生教育创新计划基金项目(JGY2014062)

摘　　要：为解决在大规模网络流量上使用SVM(support vector machine)算法时训练速度慢的问题,提出一种基于并行DAGSVM(directed acyclic graph support vector machine)的网络流量分类方法。利用有向无环图,将使用Spark中并行二分类SVM算法训练得到的子分类器组合得到并行多分类SVM分类器。对比实验结果表明,与单机SVM方法相比,该方法在确保较高分类精度的前提下,训练速度提高了100倍以上。To solve the problem of slow training speed of SVM algorithm on large network traffic dataset,a network traffic classification method based on parallel DAGSVM was proposed.The directed acyclic graph was used to combine the sub classifiers that were trained using the parallel binary SVM algorithm on Spark into parallel multi-classification SVM classifier.Compared with the single SVM method,experimental results show that the parallel method ensures high classification accuracy and improves the training speed by more than 100 times.

关 键 词：网络流量分类 分布式计算 有向无环图 支持向量机 并行 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]