Evaluating machine learning and deep learning approaches in IoT intrusion detection systems: A comparative study of techniques and datasets
Gaurav Kadam, Aditya Bansal, Yashi Tiwari, Avneesh Sinha, Rupali Pandey, Vishal Sharma · 2025
The Internet of Things (IoT) is a collection of interconnected devices that can communicate and send knowledge between them. Intrusion detection in IoT means closely monitoring network traffic to detect malicious activity. Hence, it is crucial to create a reliable Intrusion Detection System. This paper analyses different approaches used in an Intrusion Detection System that uses Machine Learning and Deep Learning algorithms. It explores various datasets built in different environments containing different attack types, providing the models with a broader perspective on which to be trained. It is discovered that the accuracy of most models was very promising; however, there are many instances where these Intrusion Detection Systems fail. Therefore, it sheds light on one such problem, false positive results, and explains ways to tackle it. It concludes with future research scope in this field, demonstrating the need for better approaches to combat new threats and make the system more trustworthy and feasible.