Machine Learning-enabled Intrusion Identification Model for IoT Environment

Abstract

The Internet of Things (IoT) encompasses millions of autonomous and diverse intelligent devices that interact with
one another without people’s involvement. IoT-based solutions have enhanced experiences across many industries,
including medical care, farming, supply chain management, learning and traffic monitoring, and utilities. Node
heterogeneity has produced security concerns, constituting one of the most complex IoT issues. Employing security
mechanisms such as data encryption, access management, and identification for IoT gadgets is unsuccessful in
ensuring safety. This research identifies different types of IoT assaults. It discusses both superficial learning techniques
and Machine Learning (ML) methods utilized in systems for Intrusion Detecting Systems (IDS) within the context of
IoT. The efficacy of these models has been assessed utilizing five benchmark databases. Performance criteria were
employed to analyze the effectiveness of shallow and ML-oriented IDS. ML-IDS surpasses shallow computer learning
in identifying IoT assaults.