基于SVR的桡动脉血压预测

Radial artery blood pressure prediction based on SVR

目的现有的电子血压计主要使用示波法及示波改进法等方法进行血压判定。当患者体征不同时,其脉搏波形也各不相同,可能无法找到脉搏波形中对应的血压值点。本文拟根据临床采集到的桡动脉脉搏波数据使用支持向量回归(support vector regression,SVR)的方法完成与之相对应的血压的预测。方法使用欧姆龙血压计采集到的6581个任意波形的血压样本作为实验数据,其中约3/4样本作为训练集。SVR利用血压样本统计信息进行血压预测。结果实验中95%以上测试样本的收缩压和舒张压误差在[-3%,3%]以内,这表明SVR能够很好地对各类波形样本对应的血压值进行预测。结论使用SVR进行任意波形脉搏波的血压预测是可行且有效的,可以用于桡动脉处血压值的预测。

Objective Oscillometric method and its improved methods are the main methods of the existing electronic sphygrnomanometer to determine the blood pressure. The pulse waveforms are different when patients have different physical conditions. The above methods cannot always able to find the corresponding pressure value. This paper intends to use the pulse wave data collected from the clinical to complete the relative blood value prediction based on the support vector regression （SVR） method. Methods We chose 6581 clinical blood samples collected by Omron electronic sphygmomanometer as experimental data, of which about 3/4 samples used as training set. Their waveforms are arbitrary. SVR uses the statistical information of blood samples to predict blood pressure. Results For more than 95% of the test samples, the error range of blood pressure is [ - 3%, 3%]. The result shows the effectiveness of the method. Conclusions Experimental results show the feasibleness and effectiveness of SVR for radial artery blood pressure prediction.