Dual-Mode Industrial IoT Predictive Maintenance

This study focuses on Industrial Internet of Things (IIoT) sensor datasets in the context of Industry 5.0, aiming to conduct in-depth analysis and mining of sensor data to systematically explore data-driven modeling and analysis methods systematically, thereby achieving classification prediction of equipment failures and regression prediction analysis of remaining useful life (RUL). The study innovatively introduced SHAP values for model interpretability analysis, revealing key premonitory features of failures and providing transparent evidence for industrial decision-making. The research results indicate that a predictive maintenance system based on a dual-modal machine learning framework can effectively simulate real industrial scenarios, providing methodological references and technical implementation pathways for building intelligent and sustainable production systems in the Industrial 5.0 era.