Voting-based Ensemble Learning approach for Cyber Attacks Detection in Industrial Internet of Things

Abstract

The Industrial Internet of Things (IIoT) is a fast developing interconnection of smart technologies that have revolutionised the way the industries work by combining sensors and devices by automating their regular operations through Internet of Things (IoT). In various application domains, IoT devices provide seamless connectivity and diversity. Because of their round-the-clock connectivity these systems and their communication channel are subjected to targeted cyber-attacks. As a result, a multi-level security solution is required to protect the industrial system. An Intrusion Detection System (IDS) can be used to counteract these cyber-attacks by analyzing the data packets for any targeted attacks in the IIoT environment. This paper proposes a cyber-attacks detection framework for IIoT by using the Voting-based Ensemble Learning approach. In the proposed framework, an ensemble of the latest and classical Machine Learning (ML) techniques like Histogram Gradient Boosting (HGB), CatBoost, and Random Forest (RF), and a hard voting classifier are employed for efficient detection of cyber-attacks. In our proposed model, CatBoost exceeds the other two with an accuracy of 99.85 %, while HGB and RF have an accuracy of 97.90 % and 98.83 % respectively.