Machine Learning Approaches for Intrusion Detection in IoT: Literature Review

The proliferation of Internet of Things (IoT) devices has revolutionized various sectors, but it has also introduced significant cybersecurity challenges. Traditional intrusion detection methods often fall short in identifying the complex and varied attacks targeting IoT networks. This chapter explores the pivotal role of machine learning (ML) and deep learning (DL) techniques in addressing these challenges. It reviews 50 studies that utilize ML for intrusion detection in IoT, focusing on the datasets, preprocessing techniques, and algorithms that have shown promise in this field. The chapter highlights the importance of selecting high-quality datasets that accurately represent IoT traffic behaviors, such as UNSW-NB15, Bot-IoT, and IoTID20. It also delves into the critical steps of data preprocessing, including feature selection and handling imbalanced data, which are essential for improving model performance. The review covers a range of ML algorithms, from Random Forest and Decision Trees to Support Vector Machines and K-Nearest Neighbors, providing insights into their effectiveness and computational requirements. By addressing these key topics, the chapter offers a comprehensive guide for enhancing intrusion detection in IoT networks, ensuring that readers gain a deep understanding of the current state-of-the-art and future directions in this critical area of cybersecurity.