Machine learning Algorithm for Non-invasive Blood Pressure Estimation Using PPG Signals

Abstract

In this study, we propose a blood pressure estimation algorithm that employs a gradient boosting regressor. A Photoplethysmography obtained from the MIMIC II database is uniformly divided to accurately estimate blood pressure. Blood pressure is estimated by extracting the features from these data. The performance of the algorithm is evaluated by analyzing R2, MSE, MAE, and time. The MSE of SBP is 7.07 mmHg, MAE is 4.33 mmHg, and $R^{2}$ is 0.58. In addition, the MSE of the DBP is 4.18 mmHg, MAE is 2.54 mmHg, and the $R^{2}$ is 0.87. This study confirmed the possibility of developing an algorithm that can accurately estimate blood pressure.