基于PPG信号的极简特征回归树血压估计模型设计
Regression tree model for blood pressure estimation using the minimalist characteristics of photoplethysmography signal
李勋 1刘丽荣 1李浩 2杨怜琳 3王志敏 3邹梅1
作者信息
- 1. 昆明学院物理科学与技术学院,云南昆明 650214
- 2. 昆明学院医学院,云南昆明 650214
- 3. 昆明市延安医院,云南昆明 650051
- 折叠
摘要
目的:提出一种基于光电容积脉搏波(PPG)的极简特征回归树血压估计模型.方法:从单路PPG信号中提取15个特征参数,利用斯皮尔曼相关系数筛选与血压相关性最高的4个参数,构建极简特征回归树血压模型.结果:极简特征回归树血压模型收缩压和舒张压的估计误差分别达到(-0.02±3.63)mmHg和(-0.04±2.10)mmHg.结论:提出的极简特征回归树血压模型结构简洁、准确率较高,这一发现对于在可穿戴设备中使用单路PPG信号进行血压估计具有重要意义.
Abstract
Objective To propose a regression tree model for the estimation of blood pressure using the minimalist characteristics of photoplethysmography(PPG)signals.Methods Fifteen characteristic parameters were extracted from the PPG signals,and the 4 parameters with the highest correlations with blood pressure were screened using the Spearman correlation coefficient to construct a regression tree model for blood pressure estimation using the minimalist characteristics.Results The estimation errors of systolic and diastolic blood pressures in the constructed model were(-0.02±3.63)mmHg and(-0.04±2.10)mmHg,respectively.Conclusion The proposed regression tree model has a simple structure and high accuracy,which is of great significance for using a single-channel PPG signal for blood pressure estimation in wearable devices.
关键词
光电容积脉搏波/极简特征/斯皮尔曼相关系数/血压估计模型Key words
photoplethysmography/minimalist characteristics/Spearman correlation coefficient/blood pressure estimation model引用本文复制引用
基金项目
昆明学院引进人才科研项目(XJ20210003)
云南省科技厅基础研究项目(ZX20240033)
出版年
2024
NETL
NSTL
万方数据