作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2011, Vol. 37 ›› Issue (6): 104-106. doi: 10.3969/j.issn.1000-3428.2011.06.036

• 网络与通信 • 上一篇    下一篇

基于SSOM的网络流量分类方法

胡 婷 a,王 勇 b,陶晓玲 b   

  1. (桂林电子科技大学a. 计算机科学与工程学院;b.信息与通信学院,广西 桂林 541004)
  • 出版日期:2011-03-20 发布日期:2011-03-29
  • 作者简介:胡 婷(1986-),女,硕士研究生,主研方向:网络安全;王 勇,教授;陶晓玲,工程师
  • 基金资助:
    国家自然科学基金资助项目(60872022);广西研究生创新基金资助项目(2010105950812M21)

Network Traffic Classification Method Based on SSOM

HU Ting a, WANG Yong b, TAO Xiao-ling b   

  1. (a. College of Computer Science and Engineering; b. College of Information and Communication, Guilin University of Electronic Technology, Guilin 541004, China)
  • Online:2011-03-20 Published:2011-03-29

摘要: 针对目前基于端口号匹配和特征码识别的流量分类方法准确率低、应用范围受限等问题,提出一种基于有监督的自组织映射(SSOM)的网络流量分类方法。该方法使用已标注类别的网络流量训练集,通过改变自组织映射(SOM)训练过程中的权值调整规则,使输出层中获胜神经元的选择更容易,各类别之间划分更清晰,从而提高分类性能。实验结果表明,SSOM的分辨率及拓扑连续性均优于SOM,对网络流量分类具有更高的准确率。

关键词: 自组织映射, 网络流量, 分类

Abstract: In order to solve the problems in current work that relies on well known TCP or UDP port numbers or interpreting the contents of packet payloads, such as low accuracy and limited application region. This paper proposes a method based on a Supervised Self-Organizing Maps (SSOM) network classification method for traffic classification. The method uses training dataset of network traffic which has been labeled the traffic classes, and changes the adaptation rule of weighs in SOM training process, which easier to choose the winning neuron of output layer and divides each class more clearly, and improves the performance of classification. Experimental results show the SSOM network algorithm has a better resolution and a more continuous mapping than SOM. Its applying to network traffic classification has a higher accuracy rate.

Key words: Self-Organizing Maps(SOM), network traffic, classification

中图分类号: