摘要
文章概述了网络流量识别技术的基本原理和发展现状,通过分析和比较基于端口映射、基于有效负载、基于行为特征和基于机器学习的4种网络流量识别方法,得出基于机器学习的流量识别方法更加适用于电力信息通信网,并着重分析了两种基于机器学习的流量识别方法:C4.5决策树算法和神经网络算法。分析结果表明:C4.5决策树算法和神经网络算法都能有效地进行网络流量识别。
This paper summarizes the basic principle and development status of network flow classification techniques.By analyzing and comparing four types of network traffic classification methods which are method based on port mapping,method based on payload,method based on the behavioral characteristics and method based on machine learning,it draws the conclusion that flow classification method based on machine learning is more applicable to electric power information communication network.In the end,this paper emphatically analyzes the two kinds of traffic classification method based on machine learning:C4.5decision tree algorithm and neural network algorithm.
出处
《信息化研究》
2015年第1期10-14,18,共6页
INFORMATIZATION RESEARCH
基金
国家电网公司2014年科技项目"电力信息通信网流量预测与管道智能化关键技术研究及应用"
关键词
电力信息通信网
网络流量识别
机器学习
C4.5算法
神经网络算法
electricity grid information and communication network
network traffic identification
machine learning
C4.5 algorithm
neural network
