Xids An Enhanced Honeynet Enabled Intrusion Detection for Industrial Cyber Physical System with Explainable Ai and Optimised Deep Learning Techniques

Abstract

In recent years, industrial cyber-physical systems (ICPS) have faced a significant newlinechallenge in maintaining robust security against a growing number of cyber threats. Traditional newlineintrusion detection systems often lack transparency, interpretability, and efficient feature newlineselection techniques, leading to suboptimal detection accuracy and a limited understanding of newlinethe reasoning behind detection decisions. Additionally, the optimisation of deep learning newlinemodels used in ICPS intrusion detection remains a complex task, requiring effective newlinehyperparameter tuning to achieve optimal performance. Strong breach detection systems are newlineneeded to make industrial cyber-physical systems (ICPS) safer because cyber dangers are newlinebecoming more common. This research presents an enhanced intrusion detection system for newlineindustrial cyber-physical systems (ICPS) called XIDS utilizing explainable artificial newlineintelligence (XAI), ensemble-based filter feature selection techniques (EFFS), enhanced Krill newlineherd optimisation (EKHO), and Bayesian optimisation algorithms to fine-tune deep-learning newlinehyperparameters newline