多元统计学方法在中医证候分类识别中的应用及展望-中国中医药信息.PDF

·124· Chinese Journal of Information on TCM Aug.2015 Vol.22 No.8 ·综述 · 多元统计学方法在中医证候分类识别中的应用及展望 徐玮斐,刘国萍,王忆勤 上海中医药大学基础医学院,上海 201203 摘要：中医证候标准化、客观化、规范化研究,是现今继承发展中医的重点、难点和热点问题。随着计算 机技术发展和多学科融合交叉的广泛应用,探索中医证候分类识别研究的方法层出不穷。本文就多元统计学方 法在中医证候分类识别中的应用现状予以分析,总结了聚类分析、主成分分析和因子分析、判别分析和 Logistic 回归分析及结构方程模型在证候分类识别中的应用现状及存在的问题,为进一步研究中医证候分类 识别方法提供一定参考。 关键词：多元统计方法；中医；证候；综述 DOI：10.3969/j.issn.1005-5304.2015.08.039 中图分类号：R2-05；R228 文献标识码：A 文章编号：1005-5304(2015)08-0124-05 Application and Prospect of Multivariate Statistical Analysis in TCM Syndrome Classification XU Wei-fei, LIU Guo-ping, WANG Yi-qin (Shanghai University of Traditional Chinese Medicine, Shanghai 201203, China) Abstract ：Standardization, objectification, and normalization of TCM syndromes are emphasis, difficulty, and hot issue of inheritance and development of TCM. With the technological development of computer and extensive application of multi-disciplinary integration, research methods for the identification of TCM syndrome classification emerge in endlessly. This article analyzed the application status of multivariate statistical analysis in the identification of TCM symdrome classification, summarized application status and existing problems of cluster analysis, PCA and factor analysis, discriminant analysis and Logistic regression analysis, and structure equation model in the identification of TCM symdrome classification, which provided references for further studies on identification methods of TCM symdrome classification. Key words ：multivariate statistical analysis；TCM ；syndrome；review 中医的精髓是辨证论治。辨证,是将四诊收集到 探索层出不穷。其中,多元统计学方法凭借其长期经 的资料、症状和体征,通过分析、综合,辨清疾病的原 验、相对成熟的算法和容易解释的统计结果,被广泛 因、性质、部位及邪正关系,将之概括、判断为某种 用于中医证候分类的研究中。