Development of Machine Learning Framework for the Protection of IoT Devices

Abstract

Internet of Things (IoT) has a wide range of threats to businesses, according to security experts. Organizations need an intelligent system that can automatically detect suspicious IoT devices linked to their networks. This study introduces a unique security framework powered by machine learning (ML) that automatically adapts to the growing security needs of the IoT sector. There should be a way to identify IoT devices that aren't on a trusted white list. In this article, a machine learning method has been used to recognize IoT device types from a white list by using network traffic data. Seventeen separate IoT devices, each representing one of nine different categories of IoT devices, were manually tagged to train and assess multi-class classifiers. The majority rule was used to classify block listed devices accurately using unidentified in 86% of trial forms, while authorized expedient categories stayed appropriately identified through the real kinds with 88% of forms. The detection times varied for different types of IoT devices. In addition, it shows how the machine learning-based IoT white-listing system can defend itself against hostile attacks.