Advancements in IoT Anomaly Detection: Leveraging Machine Learning for Enhanced Security

DOI

10.2991/978-94-6463-700-7_30How to use a DOI?

Keywords

IoT anomaly detection; machine learning; cybersecurity; ensemble learning; deep learning; class imbalance; SMOTE; real-time detection; explainable AI; adaptive models; IoT security; data preprocessing; feature extraction; lightweight models; hybrid detection methods

Abstract

The internet of things(IoT) is one of the fastest-growing technologies that has disrupted industries by allowing devices to connect without disruption. Nevertheless, the adoption of IoT devices has posed new security risks as it is hard to distinguish between normal and anomalous behaviors indicative of cyber-attacks or system issues. This paper aims at discussing the present progress of IoT anomaly detection; this has the argument that most of the solution relies on machine learning (ML). ML has been shown to be very influential in this field thanks to its ability to unravel sophisticated patterns from large, diverse datasets, something which many conventional approaches are incapable of. This paper investigates various approaches used in anomaly detection in IoT using ensemble learning, deep learning, and a combination of both. An insight into their strengths and weaknesses is given. Important directions for further research are the use of ensemble learning methods that combine several classifiers for better detection ability and deep learning models such as CNN and LSTM for temporal and spatial data analysis. A discussion also ensues to show how to overcome class imbalance in IoT datasets using SMOTE method. Also, it highlights that there is a demand for immediate dynamic anomaly detection for IoT systems since the nature of IoT is unsteady and emphasizes the utilization of the explainable AI (XAI) to enhance the users’ trust in models. We also find gaps in the literature: first, insufficient exploration of the use of ensemble methods; second, high demand for comparing different approaches; third, the necessity to develop low-power models for IoT devices. This is an area that should receive further attention when it comes to prospective work because creating better, more flexible, and finer weight structures to protect IoT network appearances against new threats should be the goal of future work. Based on the findings in this paper, there is need for continual research in the IoT anomaly detection where such concerns as performance, flexibility and interpretability need to be met to establish secure IoT systems.

Copyright

© 2025 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Rajesh Rajaan
AU  - Baldev Singh
AU  - Nilam Choudhary
PY  - 2025
DA  - 2025/04/19
TI  - Advancements in IoT Anomaly Detection: Leveraging Machine Learning for Enhanced Security
BT  - Proceedings of the International Conference on Advancements in Computing Technologies and Artificial Intelligence (COMPUTATIA-2025)
PB  - Atlantis Press
SP  - 367
EP  - 389
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6463-700-7_30
DO  - 10.2991/978-94-6463-700-7_30
ID  - Rajaan2025
ER  -