基于Shannon熵法的慢性心衰证候要素诊治规律研究

目的：对慢性心衰中医和中西医结合的临床文献进行分析,并结合数据挖掘方法,探讨该病证候要素的诊治规律。方法：以中国学术期刊全文数据库（China National Knowledge Internet,CNKI）和中国生物医学文献数据库（Chinese Biomedical Literature Database,CBM）为数据源,以＂慢性心力衰竭＂或＂慢性心衰＂或＂慢性充血性心力衰竭＂或＂慢性心功能不全＂,＂中医＂或＂证候＂或＂证＂或＂中西医＂为检索词,分别进行一次和二次检索,对文献进行筛选。数据预处理后,采用Shannon熵非线性复杂系统划分法对中医四诊和常用中药数据进行特征提取。结果：阳虚、气虚、血瘀、水停、阴虚、痰浊为慢性心衰的常见证候要素,补气药、活血药、利水药、化痰药、解表药、温里药为临床主要用药。我们提取出了与证候要素密切相关的症状和中药,通过关联系数值对各症的诊断贡献度进行了量化,各要素下分值较高的症状基本为该证的共性症状,分值较低的症状多为某一类患者或疾病本身的特异表现;同时筛选出各证候要素对应的常用中药,间接反映了该药的临床对证使用情况。结论：本研究采用Shannon熵法探索中医四诊-证候要素-方药诊治规律,为该病证候要素属性的判别和临床辨证候要素的用药选择提供了一定的理论依据。

Diagnosis and treatment rule of traditional Chinese medicine for syndrome factors of chronic congestive heart failure: a study based on Shannon entropy method

Objective： The primary aim of this research is to systematically sort out and analyze available clinical documents for chronic congestive heart failure in traditional Chinese medicine（TCM） and integrated traditional Chinese and Western medicine,and to explore the diagnosis and treatment rule of TCM syndrome factors with data mining method.Methods：Based on the China National Knowledge Internet（CNKI） and Chinese Biomedical Literature Database（CBM）,using ＂chronic heart failure＂ or ＂chronic congestive heart failure＂ or ＂chronic cardiac insufficiency＂ and ＂traditional Chinese medicine＂ or ＂syndrome＂ or ＂integrated traditional Chinese and Western medicine＂ as keywords for literature selection, we used Shannon entropy nonlinear complex system method for the feature extraction after data preprocessing.Results：The statistics showed that yang deficiency, qi deficiency, blood stasis, water retention, yin deficiency,and turbid phlegm were principal TCM syndrome factors; herbs of qi-tonifying, blood-activating, water- draining, phlegm-dispelling, exterior-releasing, interior-warming were main drugs for clinical application. The symptoms and Chinese herbs which intimately connect with TCM syndrome factors were extracted, while their diagnostic contribution degrees were quantified by correlation coefficient. The symptoms which had high scores were general syndromes while the symptoms which had low scores were specific manifestations of a certain patient. Common Chinese herbs corresponding to each syndrome factor were picked out, indirectly reflecting their frequencies in clinical application.Conclusion：Based on data mining of the relationship among the four diagnostic methods of TCM, syndrome factors and herbs, this study provided references for differentiation of TCM syndrome and selection of Chinese medicine according to clinical syndrome factors.