Intrusion Detection System Using Data Stream Classification

Amer Abdulmajeed  Abdualrahman; Mahmood Khalel  Ibrahem

doi:10.24996/ijs.2021.62.1.30

Authors

Amer Abdulmajeed Abdualrahman Informatics Institute for Post Graduate Studies, College of Science, University of Baghdad, Baghdad, Iraq
Mahmood Khalel Ibrahem College of Information Engineering, Al-Nahrain University

DOI:

https://doi.org/10.24996/ijs.2021.62.1.30

Keywords:

Intrusion Detection System, Data Stream Classification, CICIDS 2017 dataset, Feature selection

Abstract

Secure data communication across networks is always threatened with intrusion and abuse. Network Intrusion Detection System (IDS) is a valuable tool for in-depth defense of computer networks. Most research and applications in the field of intrusion detection systems was built based on analysing the several datasets that contain the attacks types using the classification of batch learning machine. The present study presents the intrusion detection system based on Data Stream Classification. Several data stream algorithms were applied on CICIDS2017 datasets which contain several new types of attacks. The results were evaluated to choose the best algorithm that satisfies high accuracy and low computation time.