Towards Stream Based Intrusion Detection System

Abstract

This research paper showcases how machine learning can be effectively used in order to increase efficiency in stream-based approach which aims at detecting intrusion in real-time. Hence increasing efficiency in real-time intrusion detection using stream-based approach which is suitable for high-speed network. It uses four classifier namely Naïve Bayes, Random forest, Decision tree and KSVM. A comparative study of all the four classifier is done and the best one is chosen for finding maximum efficiency which can be obtained when stream-based approach is used along which machine-learning.