A Hybrid Ensemble Model using XGBoost and AdaBoost to detect and distinguish zero-day attacks

Edakkat Parambil, Ajay Krishna (2023) A Hybrid Ensemble Model using XGBoost and AdaBoost to detect and distinguish zero-day attacks. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (999kB) | Preview

Abstract

This research investigates the crucial role of Intrusion Detection Systems (IDS) in addressing cyber threats, with a specific emphasis on the detection of Zero-Day assaults. Zero-day attacks, exploiting vulnerabilities concealed from developers and security experts, present a substantial security threat due to the unavailability of immediate patches. Traditionally, zero-day attack detection relied on machine learning algorithms like AdaBoost, which, while highly effective in aggregating weak learner predictions, is inefficient when dealing with complex, multi-class datasets frequently encountered in network traffic analysis. Which results in inadequate protection of critical assets, intellectual property, and sensitive data. The performance of various machine learning models is evaluated, to determine the most effective model for network intrusion detection, and to emphasise the relevance of flexibility and precision in identifying developing threats. Four unique machine learning models the AdaBoost Classifier, XGBoost Classifier, Random Forest Classifier, and a Hybrid Ensemble Model leveraging the capabilities of the above machine learning techniques for an Intrusion Detection System (IDS) that effectively identifies zero-day attacks and false positive reduction is introduced. Using a large dataset, these models are tested for their capacity to identify various network activity and, more importantly, their ability to detect Zero Day attacks. The results reveal that the Hybrid Ensemble Model achieves the highest accuracy of 82%, compared to the AdaBoost Classifier with 41%, XGBoost Classifier with 75% and Random Forest Classifier with 77%.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Sahni, Vikas UNSPECIFIED
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > QA Mathematics > Computer software > Computer Security T Technology > T Technology (General) > Information Technology > Computer software > Computer Security Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Cyber Security
Depositing User:	Ciara O'Brien
Date Deposited:	16 Apr 2025 13:58
Last Modified:	16 Apr 2025 13:58
URI:	https://norma.ncirl.ie/id/eprint/7432

Actions (login required)

View Item