Intelligent predictive maintenance: multivariate ML model optimization in an Edge-Fog-Cloud environment

Preeti Jain, Koustav Kumar Mondal, Debasis Das, Arpit Khandelwal

Abstract: As a cornerstone of the global economy, the automotive industry requires intelligent and reliable maintenance approaches. Predictive maintenance provides a proactive mechanism for timely fault detection, minimized operational interruptions, and prolonged vehicle lifespan. Traditional cloud-based predictive maintenance solutions, however, face latency, bandwidth constraints, and limited scalability challenges in real-time applications. This paper proposes FogBayes, an intelligent predictive maintenance system built on a three-layer Edge–Fog–Cloud architecture to balance computation and minimize latency, employing optimized multivariate machine learning models at its core. The proposed system model achieves a prediction accuracy of 98.88% and an AUC ( Area under curve) of 0.9830, outperforming several baseline approaches. To ensure transparency and reproducibility, the complete source code, datasets, and experimental configurations used in this study are publicly available at https://github.com/TakMashhido/FogBayes.

External IDs:dblp:journals/computing/JainMDK25