Review on Various Machine Learning Algorithms Implemented in IoT Security

Abstract

The number of devices connected using Internet of Things (IoT) technology and the data accumulated by these devices, has exploded in the last few years. Because of their resource constraints, participating devices in IoT networks can be difficult, and security on these devices is frequently disregarded. As a result, attackers now have a stronger motivation to object IoT devices. An attack on a network grows rapidly, conventional intrusion detection systems (IDS) struggle to keep up. In this study, we discuss Random Forest, K-Nearest Neighbor, Artificial Neural Network, Decision Tree, Support Vector Machine, and Naive Bayes. The Bot-IoT dataset is used to compare machine learning methods for multi-class and binary classification. We theoretically analyze the said ML algorithms based on their criteria by defining each and every algorithm.