A Random Tree Based Algorithm for Blood Pressure Estimation

Abstract

Machine learning algorithms shown great potential in medical applications. This paper shows the use of it for noninvasive estimation of blood pressure through systems based on the unsupervised collection of signals. Our work is based on the study of morphology and timing characteristics of ECG and PPG signals like Pulse Transit Time, Heart Rate and others showed in this paper. We implement Random Forest regression algorithm to reach the final result of cuff-less Blood Pressure (BP) estimation with RMS error of 13 mmHg for SBP and 12.89 mmHg for DBP.