体检人群肝脂肪病变者中医脉象信号的递归定量识别与分析

Recurrence Quantification Identification and Analysis of Traditional Chinese Medicine Pulse Signals in the Health Check-up Population with Hepatic Steatosis

目的运用研究非线性动力学的递归定量分析(recurrence quantification analysis,RQA)方法对体检人群肝脂肪病变者的脉象信号进行分析,探讨脉象信号非线性动力学特征对肝脂肪病变的识别价值。方法运用ZY-Ⅰ型脉诊仪采集体检人群的脉象信号,根据腹部超声报告将体检人群分为肝脂肪病变组和非肝脂肪病变组;提取体检人群脉象信号RQA特征,并运用非参数检验分析两组人群脉象信号的RQA特征差异;基于脉象信号RQA特征,运用随机森林机器学习方法建立体检人群肝脂肪病的识别模型,并通过评价指标准确率、精确率、召回率、F1值、受试者工作特征曲线(receiver operating characteristic curve,ROC)及曲线下面积(area under the curve,AUC)评估模型识别性能。结果肝脂肪病变组脉象信号RQA特征递归率、确定性、对角线长度的均值、递归熵、层状度、竖直/水平线段长度均值和最长竖直/水平线段长度均高于非肝脂肪病变组(P<0.05);基于脉象信号RQA特征建立的体检人群肝脂肪病变识别模型,其准确率为80.34%,精确率为82.17%,召回率为86.00%,F1值为84.04%,AUC为86.77%。结论与非肝脂肪病变组相比,肝脂肪病变组的脉象信号系统表现出更高的规律性、确定性、稳定性,基于RQA特征建立的体检人群肝脂肪病变识别模型能较好地区分肝脂肪病变组与非病变组的脉象信号,可为肝脂肪病变的早期预警及辅助诊断提供一定的临床参考。

Objective To analyze the pulse signals of the health check-up population with hepatic steatosisusing the recurrence quantification analysis(RQA)method for studying nonlinear dynamics,and to investigate the value of the nonlinear dynamical characteristics of pulse signals in identifying hepatic steatosis.Methods The ZY-Ⅰpulse diagnostic instrument was used to collect the pulse signals of the health check-up population,and the population was divided into hepatic steatosis group and non-hepatic steatosis group according to the abdominal ultrasound findings.The RQA characteristics of the pulse signals were extracted,and a non-parametric test was used to investigate the difference in RQA characteristics between the two groups.Based on the RQA characteristics of the pulse signals,the random forest(RF)algorithm was used to establish an identification model for hepatic steatosis in the health check-up population,and the performance of this identification model was evaluated based on the criteria including accuracy,precision,recallrate,F1-score,receiver operating characteristic(ROC)curve,and area under the ROC curve(AUC).Results Compared with the non-hepatic steatosis group,the hepatic steatosis group had significantly higher RR,DET,L,ENTR,LAM,TT,and V_(max) among the RQA characteristics of pulse signals(P<0.05).The identification model for hepatic steatosis in the health check-up population based on the RQA characteristics of pulse signals had an accuracy of 80.34%,a precision of 82.17%,a recall rate of 86.00%,an F1-score of 84.04%,and an AUC of 86.77%.Conclusion Compared with the non-hepatic steatosis group,the hepatic steatosis group shows higher levels of regularity,certainty,and stability in the pulse signal system.The identification model for hepatic steatosis in the health check-up population based on RQA characteristics can better distinguish the pulse signals of the hepatic steatosis group from those of the non-hepatic steatosis group,which can provide a reference for the early warning and auxiliary diagnosis of hepatic steatosis.