Intrusion Detection System Based on Machine Learning Techniques: A Survey

Abstract

The volume of network traffic data has become so big and complicated as a result of the development in Internet-based services that it is extremely difficult to process using typical data processing techniques. Due to the enormous and complicated nature of network traffic data, fast and effective cybersecurity intrusion detection is a highly difficult task. In order to detect hostile traffic as soon as feasible, a realistic cyber security intrusion detection system should be able to handle a huge volume of network traffic data as quickly as possible.This paper studies a classification-based intrusion detection algorithms based on Machine Learning for speedy and effective intrusion detection in huge network traffic, including Support Vector Machines (SVM), Random Forests(RF), Decision Trees(DT), Naïve Bayes (NB), Deep Neural Network (DNN), Extreme Gradient Boosting (XG Boost), Nondominated Sorting Genetic Algorithm (NSGA2), as well as Deep Belief Network (DBN) and artificial neural network(AI). It has been applied to ACTUAL TIME The KDD dataset, KDD Cup 99 dataset, NSL-KDD dataset, CICIDS-2017 dataset, CICIDS-2018 dataset, ISCX Dataset, CICAndMal2017 dataset and UNSW-NB15 to compare performance based on rating Intrusion detection systems which are assessed in terms of training time, prediction time, accuracy.