Design of a machine learning-based blood pressure prediction system

Abstract

The limitations of traditional cuff-type blood pressure equipment have gradually emerged due to cumbersome measurement operations and discomfort during use, which have made it difficult to meet users' needs for convenient and comfortable blood pressure monitoring. To address these shortcomings, a cuffless blood pressure monitoring method is proposed, which utilizes electrocardiogram (ECG) and photoelectric volumetric tracing (PPG) signals to predict the user's systolic blood pressure, diastolic blood pressure, and heart rate through a machine learning algorithm of elastic network algorithm. The ECG and PPG signals are collected and processed by electrodes, and the advantages of lasso regression and ridge regression are combined to improve the accuracy and efficiency of data analysis and realize the accurate prediction of blood pressure. After experimental testing and analysis, the results show that the average absolute error of the system is within ±5 mmHg, and the average relative error of systolic blood pressure and diastolic blood pressure is 3.76% and 5.36%, respectively. The measurement accuracy is comparable to that of the traditional OMRON sphygmomanometer, but it is significantly improved in terms of portability, real-time performance, and user experience, which demonstrates a good prospect of application, promotes the development of intelligent health monitoring devices, and can provide new opportunities for the development of smart health monitoring devices in the future. health monitoring devices, and can provide new ideas and directions for the development of future smart health monitoring devices.