Abstract

Since cardiovascular disease (CVD) continues to be a major cause of death worldwide, early and precise prediction tools are essential. The results of this work point to an architecture that runs ML models in the cloud on a user’s data, forecasting the risk of CVD in real time, and is mainly useful in resource-constrained environments. It applies privacy-ensuring techniques when handling patients’ data as it is transmitted and used, and it relies on secure cloud computing for managing the information on a large scale. When the aim is to use fewer computer resources without losing accuracy, lightweight models are preferred, for instance, ensemble methods and efficient decision trees. How accurate and precise it is, the quality of its F1-score and how efficiently it runs are assessed once benchmark clinical datasets have been used to validate it. Since the suggested approach is effective, reliable and secure, it works well for mHealth apps and remote healthcare, the findings indicate.