基于支持向量机网络流量识别的研究

Research on Network Traffic Recognition Based on Support Vector Machine

随着以P2P网络为特征的网络的新应用的出现和发展,已有的网络流量识别方法不能有效、准确识别出复杂的网络流量环境.为满足网络流量的要求,本文提出了基于支持向量机(SVM)的网络流量识别方法.在对一组监督学习的公用数据集和实时流量进行应用前,先建立一组恰当的训练集和实验数据集进行实验,充分证明此法准确性高,操作简单,识别效率高,且对实时流量识别有良好的现实可行性.

With the emergence and development of new network applications characterized by P2 P networks, the existing network traffic identification methods can’t effectively and accurately identify the complex environment of network traffic. In order to meet the requirements of network traffic identification, this paper proposes a network traffic identification method based on Support Vector Machine(SVM). Before the application of a set of supervised learning public data sets and real-time traffic, a set of appropriate training sets and experimental data sets are established to carry out experiments. Finally, it is fully proved that this method has high accuracy, simple operation, high recognition efficiency, and good practical feasibility for real-time traffic identification.