A Hybrid ABNB Model for Detecting Malicious Attacks for IIoS

Industry 4.0 is no longer just a futuristic idea but is gradually becoming a reality, along with its key components, IIoT and IIoS. Industrial 4.0 introduces IIoT and IIoS technologies to improve manufacturing and production processes. IIoS and IIoT collect sensitive and proprietary data for automation and analytics, which makes industrial automated systems an attractive target for malicious attacks. Therefore, adequate safeguards and robust security controls are required for IIoS. Numerous machine learning strategies have been intensively studied to develop efficient and intelligent security measures. However, most of the current ML-based strategies experience difficulties when used on real-life operational grounds.The difficulties include uncertainty in determining the most appropriate ML approach based on an industrial model. Moreover, these challenges are aggravated by poor performance on inconsistent datasets and excessive rates of false positives. Therefore, the purpose of this study is to investigate the performance of various Machine Learning algorithms, including supervised learning (Nave Bayes), ensemble learning (AdaBoost), neural networks (MLP), and hybrid algorithm (ABNB), in detecting malicious attacks on three different datasets.