A Semantic Architecture for Continuous Health Monitoring, Risk Prediction, and Proactive Decision Making

Wearable sensors combined with health records and user-submitted data are becoming ubiquitous for continuous health monitoring. Semantic web technologies are well suited for representing and reasoning over these heterogeneous health data. However, existing semantic health monitoring architectures have notable deficiencies, particularly in interoperability, situation prediction, and uncertainty handling. We propose a semantic architecture that integrates an ontology with rules, fuzzy inference, and machine learning to detect and predict health risks using heterogeneous health data. We illustrate its application through a use case of atrial fibrillation and demonstrate its ability to detect and predict health situations, as well as provide decision support aligned with established health workflows and clinical guidelines.