Machine Learning-Based Ids for IoT: A Comparative Analysis

Abstract

Now that we can connect everyday objects to the Internet through integrated terminals, the Internet of Things (IoT) has become a fundamental part of our life and a key component in many fields like Health care, Industry, and Education.IoT refers to the network of physical devices that are embedded with sensors, and software, enabling them to collect and exchange data over the Internet.These devices can communicate with each other, share data, and perform actions based on that data, creating a seamless and automated network of connected objects.However, IoT security has come under intense scrutiny after various high-profile attacks where typical IoT devices were used to penetrate and attack the network of many important organizations.Furthermore, traditional IoT security measures, such as firewalls along with authentication and encryption protocols are no more efficient against modern and sophisticated attacks.Motivated by this fact, we decided to perform a comparative analysis of four well-known Machine Learning algorithms to identify their efficiency in classifying attacks on the newest IoT dataset called EDGE-IIOTSET 2022.In this context, the present paper illustrates and analysis the results of a comparative analysis of the four classifiers named: Support Vector Machine (SVM), Naive Bayes (NB), Decision tree (DT), and Light Gradient Boosted (Light GBM).Moreover, and to reduce dimensionality and enhance the model's performance the Pearson Correlation coefficient is used.Our empirical results of experiments conducted on the aforementioned dataset indicate that the best classification accuracy was achieved by Light GBM followed by SVM.