IntrusionHunter: Detection of Cyber Threats in Big Data

Hashem Mohamed, Alia El Bolock, Caroline Sabty

2023

Abstract

The rise of cyber-attacks has become a serious problem due to our growing reliance on technology, making it essential for both individuals and businesses to use efficient cybersecurity solutions. This work continues on previous work to improve the accuracy of intrusion detection systems by employing advanced classification techniques and an up-to-date dataset. In this work, we propose IntrusionHunter, an anomaly-based intrusion detection system operating on the CSE-CICIDS2018 dataset. IntrusionHunter classifies intrusions based on three models, each catering to different purposes: binary classification (2C), multiclass classification with 7 classes (7C), and multiclass classification with 15 classes (15C). Four main classification models were used: Random Forest, Extreme Gradient Boosting, Convolutional Neural Networks, and Deep Neural Networks. The results show that Random Forest and XGBoost algorithms outperformed state-of-the-art intrusion detection systems in binary and multiclass classification (15 classes). The findings also show that the dataset imbalance needs to be addressed to improve the performance of deep learning techniques.

Paper Citation

in Harvard Style

Mohamed H., El Bolock A. and Sabty C. (2023). IntrusionHunter: Detection of Cyber Threats in Big Data. In Proceedings of the 12th International Conference on Data Science, Technology and Applications - Volume 1: DATA; ISBN 978-989-758-664-4, SciTePress, pages 311-318. DOI: 10.5220/0012081900003541

in Bibtex Style

@conference{data23,
author={Hashem Mohamed and Alia El Bolock and Caroline Sabty},
title={IntrusionHunter: Detection of Cyber Threats in Big Data},
booktitle={Proceedings of the 12th International Conference on Data Science, Technology and Applications - Volume 1: DATA},
year={2023},
pages={311-318},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012081900003541},
isbn={978-989-758-664-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 12th International Conference on Data Science, Technology and Applications - Volume 1: DATA
TI - IntrusionHunter: Detection of Cyber Threats in Big Data
SN - 978-989-758-664-4
AU - Mohamed H.
AU - El Bolock A.
AU - Sabty C.
PY - 2023
SP - 311
EP - 318
DO - 10.5220/0012081900003541
PB - SciTePress