The Internet Traffic Classification an Online SVM Approach
Yuhai Liu, Hongbo Liu, Hongyu Zhang, Xin Luan · International Conference on Information Networking · 2008
Accurate and quick classification of Internet traffic is of fundamental importance to numerous network activities, such as quality of service, security monitoring and network management. So accurate, quick, effective classification is necessary. In this paper, we apply online SVM technique for Internet traffic identification and compare the result with that of previously applied naive Bayes kernel estimation in AUCKLAND Vi and Entry data sets. Our results show that online SVM technique is more robust and accurate than naive Bayes algorithm. The test error can be limited to 5.81% in Entry data sets. For AUCKLAND Vi data sets, the test error can be limited to 14.05% and greatly outperforms naive Bayes kernel estimation.