Advances in Ensemble Machine Learning for Network Intrusion Detection Systems: A Comprehensive Review

Salama, Mohmed A.; Tawfeek, Radwa M.; Hamdy, Sara; Salim, Omar M.

doi:10.21608/bjest.2025.445861

Advances in Ensemble Machine Learning for Network Intrusion Detection Systems: A Comprehensive Review

Document Type : Research Article

Authors

Electrical Engineering Department, Benha Faculty of Engineering, Benha University, Benha, Egypt

10.21608/bjest.2025.445861

Abstract

As cyber threats grow increasingly sophisticated, robust network security demands adaptive intrusion detection systems (IDS). Traditional machine learning-based IDS often struggle with high false alarm rates and poor generalization to emerging attacks, while deep learning-based IDS offer high detection accuracy but require significant computational resources. Ensemble learning techniques provide an effective balance between efficiency and accuracy, improving detection through model diversity and decision aggregation. This review explores ensemble-based intrusion detection systems, emphasizing diverse aggregation techniques, including homogeneous and heterogeneous ensemble methods. It provides an in-depth analysis of feature selection strategies, data balancing techniques, and classification models, offering a comparative assessment across benchmark datasets. Additionally, the study highlights key challenges and outlines future research directions to advance ensemble learning in network intrusion detection.

Keywords