Machine Learning for Hypertension Risk Prediction in Precision Public Health

Science Department, Xi’an Jiaotong-Liverpool University, 215123 Suzhou, China

^* Corresponding author: This email address is being protected from spambots. You need JavaScript enabled to view it.

Abstract

Early identification of hypertension risk is essential for cardiovascular disease prevention, but traditional logistic and Cox regression models are limited by linear assumptions and modest generalisability. Machine learning (ML) can model complex, high-dimensional risk patterns and has shown improved predictive performance. This review summarizes ML applications in hypertension prediction from three perspectives: data sources, modelling methods, and clinical applications. Electronic health records and population cohorts are the main data sources, while multi-omics and wearable data remain underused. Ensemble methods and LSTM models consistently outperform conventional approaches, and interpretability tools such as SHAP help explain predictions. Clinically, ML enables early identification of prehypertension and improves community screening efficiency. These advances highlight the growing potential of ML as a practical tool in hypertension management. Key challenges include limited external validation, insufficient multimodal integration, and concerns about fairness and implementation. Future progress depends on developing clinically trustworthy, generalisable, and practically useful models for precision public health.