Smart Industry Anomaly Detection Dataset for Cyber-Physical Threat Monitoring in Industrial IoT Environment

This project focuses on identifying cyber-physical threats in Industrial IoT (IIoT) systems by analyzing network traffic data. Utilizing advanced machine learning models like Stacking Classifier and XGBoost, the project classifies network traffic into five distinct attack categories: Normal, Probing, Denial of Service (DoS), User to Root (U2R), and Remote to Local (R2L). The dataset contains 125,973 samples with 41 features, which serves as the foundation for training and testing the models. The research aims to enhance real-time anomaly detection in IIoT environments by developing a robust monitoring system capable of identifying potential cyber-attacks, ensuring the security and resilience of industrial networks. The XGBoost model achieved an accuracy of 99.66%, while the Stacking Classifier model reached an accuracy of 99.72%, demonstrating their effectiveness in detecting cyber threats and anomalies within IIoT systems.

Keywords: Smart Industry, Anomaly Detection, Cyber-Physical Threats, Industrial IoT, XGBoost, Stacking Classifier, Real-Time Monitoring, Machine Learning, Cyber-Attack Detection, DoS, Probing, U2R, R2L.