Network Traffic Classification Analysis Using Machine Learning Algorithms

In the world of networking, it sometimes becomes essential to know what types of applications flow through the network for performance of certain tasks. Network traffic classification sees its main usage among ISP’s to analyze the characteristics required to design the network and hence affects the overall performance of a network. There are various techniques adopted to classify network protocols, such as port-based, pay-load based and Machine Learning based, all of them have their own pros and cons. Prominent nowadays is Machine Learning technique due to its vastness in usage in other fields and growing knowledge among researchers of its better accuracy among others when compared. In this paper, we compare two of the basic algorithms, Naïve Bayes and K nearest algorithm results when employed to networking data set extracted from live video feed using Wireshark software. For an implementation of Machine learning algorithm, python sklearn library is used with numpy and pandas library used as helper libraries. Finally, we observe that K nearest algorithm gives more accurate prediction than Naïve Bayes Algorithm, Decision Tree Algorithm and Support Vector Machine.