摘要
针对中医辨证分型的建模问题,提出了基于数据包络分析法的证候序列量化辨证分型算法。该方法可分为3步:首先,结合中医理论,对候选辨证证型的机理进行分析与分类,获得病机解释的基本单元,并据此建立证候集合;然后,按照证候的发生与演变过程,寻找对应方证资料的用药规律和症状表现,利用数据包络分析数学工具,计算出证候在各个阶段的状态数据以获得量化的证候序列;最后,以证候序列作为度量尺度分别对候选证型和诊断信息进行量化,通过相似度计算获得诊断信息与候选证型的匹配程度,最终完成辨证分型计算。模型推理结果表明,该算法能够承载方证资料所隐含的辨证规律,从全局把握证候的动态过程,并通过对临床案例的计算分析,获得了正确的结果,这为中医的量化建模提供了一种思路。
To raise the syndrome sequence quantification, differentiation and classification algorithm based on data envelopment analysis for solving the modeling issue of syndrome differentiation and classification of traditional Chinese medicine （TCM）. This algorithm has three steps：first, in order to obtain basic units for explaining pathogenesis, and establish a syndrome collection on this basis mechanisms of syndrome differentiation and classification were analyzed and classified according to TCM theory, mechanisms of syndrome differentiation and classification were analyzed and classified according to TCM theory; second, regularity and syndromes of corresponding prescriptions were sought according to the incidence and development progress of syndromes, and mathematical tools of data envelopment analysis were used to calculate state data of syndromes in each stage and obtain quantitative syndrome sequence; finally, syndrome sequence was taken as the measurement standard to quantify candidate syndromes and diagnostic information, and the similarity was calculated to obtain the matching degree between diagnostic information and candidate syndromes, so as to complete the syndrome differentiation and classification calculation. According to the results of model-based reasoning, the algorithm could indicate the regularity implied in prescription materials, and grasp the dynamic process of syndromes in an all-round way, and its results were verified through calculation and analysis on clinical cases. At least, it provides an idea for quantitative modeling of TCM.
出处
《中国中药杂志》
CAS
CSCD
北大核心
2013年第10期1631-1642,共12页
China Journal of Chinese Materia Medica
基金
国家"十二五"科技支撑项目(2013BAI13B06)
国家基础研究发展计划(973)项目(2009CB522701)
关键词
中医辨证分型
数据包络分析DEA
证候序列
相似度测度
syndrome differentiation and classification of traditional Chinese medicine
data envelopment analysis （DEA）
syndrome sequence
measurement of matching degree
