Machine Learning Based Intelligent System for IP Traffic Classification

Abstract

As the world has become a global village, internet is considered as preliminary component of each institution, corporation, business as well as an individual. As internet base applications varies on the basis of functionality they provide to the end users, their bandwidth requirements also varies on the basis of Quality of Service (QoS) they provide. Due to the new trend introduced for classification in internet applications like use of dynamic port numbers i.e. P2P applications, classical classification techniques i.e. port based and pay-load based classification are considered inefficient. So to minimize the risk of inefficient IP data classification researchers have adopted Machine Learning (ML) based IP data classification methods. In this paper an effort is made by introducing different classifiers to increase the efficiency of existing Machine Learning techniques for IP data classification