A Robust Intrusion Detection System using Ensemble Machine Learning

In today's world there is a rising need for network attack analysis because of rising cyber threats and attacks worldwide. Network traffic, if monitored dynamically in real-time could prevent a big cyber attack or even alert before. In this paper, UNSW-NB 15 dataset is used on Ensemble method to analyze the network traffic. The major contribution of this paper is the novel Algorithms powered by boosting algorithm to come up with the best classifier from the list of classifiers. It compares the classifier in terms of accuracy as well as training time which is significant in real-time analysis of network traffic. We performed experiment in which we took 10 classifiers and our proposed algorithm came up with XGB Classifier as the best one from the set in terms of accuracy and training time combined. We have demonstrated the comparison between running time of complete experiment on Central Processing Unit (CPU) and Graphical Processing Unit GPU).