Hybrid Classification Model for Network Traffic Classification using SMOTE and PCA

Network Traffic classification (NTC) involves identifying diverse kinds of applications or traffic data for which the received data packets are analyzed. It is an important technique in communication networks, in recent times. The process to classify the network traffic classification is executed such as to take the data for input, pre-process the data, extract the attributes, classification, and performance analysis. Classifying the network traffic is a crucial process to manage the network and rapid advances in machine learning have driven the utility of learning methods to categorize traffic across a network. This work introduces a new framework to classify the traffic in network so that the efficacy is enhanced. The proposed model is the combination of SMOTE which solve problem of class unbalancing, PCA for the feature reduction and for the classification three classifiers are merged together which are SVM, KNN and random forest. The performance of proposed model is tested in Python is executed using Anaconda to test the efficacy of the introduced framework.