Intrusion detection systems for IoT based on machine learning
Amritpal Kaur, Rajesh Kumar Tyagi, Shweta Sinha, Prashant Vats · 2025
As the numeral of Internet of Things (IoT) users, amenities, and applications rises, here is an urgent necessity for a consistent and small-footprint safety solution that can be used in this setting. Security issues are more difficult to fix because cloud computing is a public system. Intrusion detection systems are one potential remedy for this issue. Because machine learning-based intrusion detection system (IDS) automatically update to guard against every new attack vector, they are growing in popularity. Owing to IDS’s significance for IoT, which is cloud-based, important publications and practices in the field are extensively researched. IDSs in cloud-based IoT fall into three primary classifications: rule-based, pattern-based, and learning processes. These findings demonstrate that precision and detection pose an enormous obstacle in IDS, which numerous academics are attempting to recover. Furthermore, the most widely used centralized or cloud-based IDS encounters problems with high latency and large network overhead when more connected objects are employed, which delays the discovery of unauthorized end users and renders the system useless against attacks. The results will be useful to academics and might inspire more research. This review’s goal is to provide an overview of studies that are pertinent to IDS for worries about IoT security. We focus on potential solutions and support for cybersecurity issues related to IoT issues.