Research on the characteristics of pulse diagnosis information in people with qi deficiency based on principal component analysis and LS / Lasso regression
FAN Xiqian1 AN Erxia1 SONG Shibo2 WANG Xingya1 LIU Luo1 LIU Jia2 WANG Nanyue2
1.School of Basic Medicine, Shaanxi University of Chinese Medicine, Shaanxi Province, Xianyang 712046, China;
2.Experimental Research Center, China Academy of Chinese Medical Sciences, Beijing 100700, China
Abstract：Objective To analyze the characteristics of pulse diagnosis information of people with qi deficiency constitution in Shaanxi Province. Methods From March 2019 to April 2020, the traditional Chinese medicine constitution questionnaire information and the information of cun, guan, and chi pulse diagnosis were collected from the medical examiners of the Second Affiliated Hospital of Shaanxi University of Chinese Medicine and students of Shaanxi University of Chinese Medicine. The constitution was classified according to the information of the traditional Chinese medicine constitution questionnaire, and the qi deficiency constitution group and the peaceful constitution group were screened out. The pulse diagnosis information of the two groups was extracted through preprocessing and characteristic parameters based on harmonic fitting, and a total of 193 characteristic parameters were obtained. Principal component analysis based on unsupervised learning and LS regression, Lasso regression, and other analysis methods based on supervised learning were used for classification and judgment, and a judgment model of pulse diagnosis information for people with qi deficiency was established, and the time domain and frequency domain characteristic parameters of pulse diagnosis information for people with qi deficiency constitution were mined. Results After screening for data integrity and pulse diagnosis signal quality, a total of 119 subjects were selected for modeling, including 61 people with qi deficiency and 58 people with peaceful constitution group. The discriminant accuracy rates of the constructed models were 76%, 72%, and 65%, respectively; the pulse number, right cun s, right foot s, right cun s2, and right foot s2 in the qi deficiency group were higher than those in the Pinghe quality group, and the energy of the Zuoguan was lower than that in the peaceful constitution group, and the differences were statistically significant (P ＜ 0.05). Conclusion The physiological differences between qi deficiency constitution and peaceful constitution are reflected in the time-frequency domain parameters of radial artery pulse diagnosis information. Through the pulse diagnosis information data model, the people of two groups can be more accurately classified and identified, which provides an objective basis for the identification of qi deficiency to a certain extent.
樊西倩1 安二匣1 宋诗博2 王星雅1 刘跞1 刘佳2 汪南玥2. 基于主成分分析和LS/Lasso回归的气虚质人群脉诊信息特征研究[J]. 中国医药导报, 2022, 19(16): 21-25.
FAN Xiqian1 AN Erxia1 SONG Shibo2 WANG Xingya1 LIU Luo1 LIU Jia2 WANG Nanyue2. Research on the characteristics of pulse diagnosis information in people with qi deficiency based on principal component analysis and LS / Lasso regression. 中国医药导报, 2022, 19(16): 21-25.
京公网安备 11010502046598号