Federated Learning Across Edge Devices for Privacy-Preserving Smart Healthcare

Abstract

The increasing reliance on data-driven technologies in healthcare has amplified privacy concerns, especially when dealing with sensitive patient data. This research proposes a federated learning (FL) approach deployed across edge devices to build intelligent healthcare systems that preserve patient privacy. By decentralizing the model training process and allowing data to remain on local devices, FL enables the use of rich datasets while maintaining confidentiality. This study employs real-world datasets, regression models, and predictive analysis to evaluate performance across various metrics such as model accuracy, latency, and data leakage risk. Results demonstrate that FL significantly improves privacy metrics and model robustness while reducing dependency on centralized cloud servers. This work contributes to the evolving paradigm of privacy-preserving AI systems in smart healthcare. Keywords Federated Learning, Edge Computing, Smart Healthcare, Privacy Preservation, Machine Learning, Predictive Analysis