CAJ | 학술논문

目的采用决策树的数据挖掘方法探讨2型糖尿病患者阳虚证与临床常规检测指标间的决策树模型.方法通过横断面调查得到249例合格2型糖尿病患者,进行证候的判断和临床指标的收集,在t检验、非参数检验、Pearson相关分析基础上,进一步采用决策树的数据挖掘方法探讨阳虚证与临床指标间的关系.结果 249例患者中有78例(31.33％)为阳虚证,以脂蛋白a、游离T3、促甲状腺激素、空腹胰岛素4个核心指标建立了阳虚证的决策树模型,10倍交叉验证得到模型的灵敏度为74.36％、特异度为89.47％,检测总正确率为84.74％.结论决策树模型可以清晰、直观的进行2型糖尿病患者阳虚证的判断,为证候客观化研究提供了契机.
목적 채용결책수적수거알굴방법탐토2형당뇨병환자양허증여림상상규검측지표간적결책수모형.방법 통과횡단면조사득도249례합격2형당뇨병환자,진행증후적판단화림상지표적수집,재t검험、비삼수검험、Pearson상관분석기출상,진일보채용결책수적수거알굴방법탐토양허증여림상지표간적관계.결과 249례환자중유78례(31.33％)위양허증,이지단백a、유리T3、촉갑상선격소、공복이도소4개핵심지표건립료양허증적결책수모형,10배교차험증득도모형적령민도위74.36％、특이도위89.47％,검측총정학솔위84.74％.결론 결책수모형가이청석、직관적진행2형당뇨병환자양허증적판단,위증후객관화연구제공료계궤.
Objective To establish the decision tree model of Yang deficiency syndrome and clinical conventional indexes in type 2 diabetes mellitus patients.Methods Syndrome decision and clinical indexes collection from 249 type 2 diabetes mellitus patient were observed and analyzed.Tree structure model were built to summarize the correspondence between Yang deficiency syndrome and clinical conventional indexes based on T test,nonparametric analysis,and Spearman correlation analysis.Results The Yang deficiency syndrome accounted for 31.33％ of 249 type 2 diabetes mellitus patients.The accuracy identification rate of tree structure model of Yang deficiency syndrome with four core index,such as LPa、FT3、TSH、FINS was 84.74％,the sensitivity and specificity were 74.36％ and 89.47％.Conclusion Decision tree model can identify Yang deficiency syndrome of type 2 diabetes mellitus patients clearly and more intuitive.Decision tree model can provide the chance of syndrome objective.