Construction of Ensemble Model for Network Traffic Classification

Abstract

Network Traffic Classification (NTC) is a key part to monitor the network, for providing the Quality-of-Service and securing the network. Machine Learning algorithms have drawn the consideration for numerous specialists during the most recent couple of years as a promising answer for NTC. Ensemble Learning is a kind of ML strategy, in which multiple base models are collaborated to fabricate more complex models as per the given training and classification techniques. The essential point here is, EL method is to integrate numerous models that are utilized for solving various tasks in order to come up with an upgraded composite global model, which is more accurate and reliable than what can be achieved with a single or base model. The generalization abilities of EL methods are stronger than that of the base models and also better to get accurate predictions. In this paper, we presented an EL method that exploits the imbalanced traffics in a network to solve the NTC problem. Although the resulting models exhibit noteworthy results in terms of accuracy contrasted with single models, with a penalty in terms of additional time and cost, these methods obstruct use of these techniques to NTC.