Prediction of hypertension risk based on multiple feature fusion.

Compared with the state-of-the-art models, the multi-feature fusion model effectively utilizes the patients' correlated multimodal features, and has higher classification accuracy and generalization performance.