Modified Hyper Parameter Optimization for enhancing intrusion detection in Industrial-Internet of Things

Abstract

The widespread deployment of Industrial Internet of Things (IIoT) devices has greatly improved operational efficiency and connectivity across various industries. However, this rapid expansion also increases network vulnerability to sophisticated cyberattacks. Traditional intrusion detection systems (IDSs) struggle with the high dimensionality and dynamic nature of IIoT data, resulting in suboptimal threat detection. To address these challenges, this work proposes a novel IDS framework combining decision trees with Bayesian hyperparameter optimization. The framework is designed to enhance detection accuracy, adaptability, and resilience in complex IIoT environments while reducing false positives. Experimental evaluation on extensive IIoT datasets demonstrates that the proposed method achieves 96.07% accuracy and a weighted average F1 score of 96.03%, outperforming conventional approaches. The results highlight the importance of strategically tuning hyperparameters and integrating decision trees with ensemble techniques to improve IDS performance. This study establishes new benchmarks for IIoT security and offers a scalable, effective solution for intrusion detection in industrial networks, providing a significant contribution to the advancement of cybersecurity in IIoT contexts.