可见光光谱用于中医证型的快速分类
Objectification of Tcm Tongue Diagnosis Based on Spectrometry
传统的中医望诊由于受到外界环境变化的干扰和医生经验及主观判断的影响，得出的结论往往存在偏差，并且肉眼的观察缺乏统一的判定标准。本文提出一种应用光谱法来反映舌部信息的方法，使用光谱仪采集舌部的光谱数据，再通过偏最小二乘和支持向量机的方法，找出光谱数据与中医证型之间的联系，最终达到将中医证型客观化的目的。实验用可见光光谱，对采集到的53例表寒里热、37例健康人作分析，分别使用支持向量机和偏最小二乘的方法对数据进行分类和预测，两种方法都可以100%正确地将两类数据区分开，但从预测精度上来讲，支持向量机的预测误差平方均值为0.07734，要优于偏最小二乘的0.13539，所以支持向量机的方法用于可见光光谱的分类要更优一些。本文提出了中医舌诊的客观化的一种全新研究思路，并为进一步进行中医症候的分类打下基础。
Interference from varying environment, experience of doctors and subjective judgment will influence the inspection of TCM, leading to deviation of diagnosis, and also lack of unified judging standard by visual inspection. This paper puts forward a method to observe the information the tongue carrying using spectrometry, collect spectrum data by instrumental means, and seek for association between spectrum data and syndrome differentiation, finally achieve the goal of externalization of application of syndrome differentiation. In the experiment, we used visible spectra and collected 53 exterior cold and interior heat patients, 37 healthy samples. For data modeling, used partial least square(PLS) and support vector machine(SVM), both of the method can predict the two sorts of data perfectly with 100%, but considering mean squared error(MSE), MSE of SVM was 0.07734 which was better than PLS's 0.13539. So SVM is a better way to modeling spectrum data.
贾萍、孙景瑞、李哲、熊慧、张晶、李刚、林凌、赵静
中医学医学研究方法基础医学
光谱学舌诊偏最小二乘法支持向量机中医症候
spectroscopytongue diagnosispartial least squaresupport vector machineTCM syndrome type
贾萍,孙景瑞,李哲,熊慧,张晶,李刚,林凌,赵静.可见光光谱用于中医证型的快速分类[EB/OL].(2011-06-10)[2025-08-11].http://www.paper.edu.cn/releasepaper/content/201106-200.点此复制
评论