Research Article Open Access

Random Forests for Online Intrusion Detection in Computer Networks

Heitor Scalco Neto1, Wilian Soares Lacerda2 and Rafael Verão Françozo1
  • 1 Instituto Federal de Mato Grosso do Sul, Brazil
  • 2 Universidade Federal de Lavras, Brazil

Abstract

This study proposes a methodology to build an Online Network Intrusion Detection System by using the Computational Intelligence technique called Random Forests and an API to preprocess the network packets. The experiments were carried out from two network traffic databases: The ISCX (i); and a test database (ii) created with the proposed API in our own network environment. The results obtained with the Random Forests technique show accuracy rates around 98%, bringing significant advances in the area of Intrusion Detection and affirming the high efficiency of the use of the technique to solve problems of intrusion detection in real network environments.

Journal of Computer Science
Volume 17 No. 10, 2021, 905-914

DOI: https://doi.org/10.3844/jcssp.2021.905.914

Submitted On: 20 April 2021 Published On: 19 October 2021

How to Cite: Neto, H. S., Lacerda, W. S. & Françozo, R. V. (2021). Random Forests for Online Intrusion Detection in Computer Networks. Journal of Computer Science, 17(10), 905-914. https://doi.org/10.3844/jcssp.2021.905.914

  • 464 Views
  • 275 Downloads
  • 0 Citations

Download

Keywords

  • Intrusion Detection Systems
  • Computer Networks
  • Computational Intelligence
  • Random Forests