Article: Performance analysis of ensemble learning classifiers for intrusion detection in IoT paradigm Journal: International Journal of Sensor Networks (IJSNET) 2026 Vol.50 No.4 pp.223 - 246 Abstract: IoT has emerged as a transformative paradigm connecting billions of smart devices, but its rapid expansion raises critical challenges such as security vulnerabilities, data breaches, and large-scale cyberattacks. Intrusion detection systems (IDS) play a vital role in mitigating these issues by monitoring network traffic and identifying malicious behaviour to enhance IoT resilience. In recent years, machine learning (ML) and ensemble learning (EL) have significantly impacted IDS by enabling adaptive and efficient detection of sophisticated threats. While ML-based approaches improve attack detection to a certain extent, EL further outperforms standalone ML models by combining multiple learners to enhance classification accuracy, robustness, and generalisation. To validate this claim, we conduct experiments on NF-UNSW-NB15v2 dataset, where results reveal that EL approaches consistently achieve superior detection performance compared to conventional ML techniques. Comparative analysis reveals that ensemble-based IDS significantly reduces false alarms while achieving higher accuracy and balanced detection rates across diverse attack categories. Inderscience Publishers - linking academia, business and industry through research

Title: Performance analysis of ensemble learning classifiers for intrusion detection in IoT paradigm

Authors: Aishwarya Vardhan; Prashant Kumar; Lalit Kumar Awasthi

Addresses: Department of Computer Science and Engineering, Dr. BR Ambedkar National Institute of Technology Jalandhar, India ' Department of Computer Science and Engineering, Dr. BR Ambedkar National Institute of Technology Jalandhar, India ' Sardar Patel University, Mandi, Himachal Pradesh, India

Abstract: IoT has emerged as a transformative paradigm connecting billions of smart devices, but its rapid expansion raises critical challenges such as security vulnerabilities, data breaches, and large-scale cyberattacks. Intrusion detection systems (IDS) play a vital role in mitigating these issues by monitoring network traffic and identifying malicious behaviour to enhance IoT resilience. In recent years, machine learning (ML) and ensemble learning (EL) have significantly impacted IDS by enabling adaptive and efficient detection of sophisticated threats. While ML-based approaches improve attack detection to a certain extent, EL further outperforms standalone ML models by combining multiple learners to enhance classification accuracy, robustness, and generalisation. To validate this claim, we conduct experiments on NF-UNSW-NB15v2 dataset, where results reveal that EL approaches consistently achieve superior detection performance compared to conventional ML techniques. Comparative analysis reveals that ensemble-based IDS significantly reduces false alarms while achieving higher accuracy and balanced detection rates across diverse attack categories.

Keywords: intrusion detection systems; IDS; internet of things; IoT; machine learning; ML; ensemble learning; EL; network-based datasets; network security; false positive alarms.

DOI: 10.1504/IJSNET.2026.153103

International Journal of Sensor Networks, 2026 Vol.50 No.4, pp.223 - 246

Received: 02 Sep 2024
Accepted: 20 Sep 2025
Published online: 22 Apr 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Performance analysis of ensemble learning classifiers for intrusion detection in IoT paradigm

Keep up-to-date