计算机工程

• 软件技术与数据库 • 上一篇    下一篇

基于相关信息的网络流量贝叶斯分类法改进

赵英,谭杨   

  1. (北京化工大学信息中心,北京 100029)
  • 收稿日期:2015-03-25 出版日期:2016-03-15 发布日期:2016-03-15
  • 作者简介:赵英(1966-),男,教授、博士,主研方向为云计算、分布式系统;谭杨,硕士研究生。
  • 基金项目:

    国家科技支撑计划基金资助项目(2010BAC67B05)。

Improving for Network Traffic Bayes Classification Method Based on Correlation Information

ZHAO Ying,TAN Yang   

  1. (Information Center,Beijing University of Chemical Technology,Beijing 100029,China)
  • Received:2015-03-25 Online:2016-03-15 Published:2016-03-15

摘要:

网络应用的快速增长导致传统基于端口与有效载荷的网络流量分类方法效率大幅降低,并且目前多数网络流分类方法未考虑流之间的相关性。针对上述问题,基于相关信息提出一种改进的网络流量贝叶斯分类法。利用流包模型反映网络流的相关信息,将非参数核密度估计方法引入贝叶斯分类器中,对分布密度函数进行非参数核密度估计。实验结果表明,与使用核密度估计和流包的分类方法相比,该方法的分类准确率更高。

关键词: 网络流量分类, 朴素贝叶斯分类, 核密度估计, 相关信息, 机器学习

Abstract:

With the rapid growth of network applications,the efficiency of traditional network traffic classification method based on ports and payloads is reduced greatly.Meanwhile,most traffic flow classification methods do not consider the correlation among the flows at present.Aiming at these problems,this paper proposes an improved network traffic Bayes classification method based on correlation information.Flow correlation information is modeled by Bag of Flow(BoF),and a method of nonparametric kernel density estimation applied to the Bayes classifier is introduced.The Bayes classification method is improved by performing the nonparametric kernel density estimation to various types of distribution functions.Experimental results show that this improved method can achieve higher classification accuracy compared with KNB and BoF-NB method.

Key words: network traffic classification, naive Bayes classification, kernel density estimation, correlation information, machine learning

中图分类号: