Anomaly Detection in IoT Networks—Classifications and Analysis Techniques

The Internet of Things is a technology that we’ve been talking about for numerous years. It defines itself as a global network consisting of interconnected services and various intelligent objects. Also, it aims to support human activities of daily life, through its sensing, computing, and communication capabilities. This network may have many risks of cyber security, so anything connected to the Internet can be exposed to cyber-attacks. Experience has shown that encryption and authentication alone aren’t enough to secure an IoT network, and detection systems are needed to detect and avoid attacks from dangerous nodes. Furthermore, designing and developing new systems based on machine learning (ML) and deep learning (DL) that can detect anomalies and attacks in IoT Networks is important. In this context, the problem of outlier detection is one of the most important problems that require more research and customized solutions. The challenge is to identify anomalies and classify them as errors that should be ignored, or as critical events that require action to prevent further service degradation. This paper aims to analyze and assess the effectiveness of anomaly detection models rooted in machine learning for IoT networks. A couple of machine learning models were compared, and their classification was also discussed for this purpose.