基于关联规则及无监督分类方法探寻孟河医派治疗中风病的用药规律  

Based on the Association Rule and Unsupervised Classification Method,the Treatment of Stroke Disease in Menghe Medical Party was Explored

作　　者：代德金 赵杨[1] 吴天晨 杨卉[2] DAI Dejin;ZHAO Yang;WU Tianchen;YANG Hui(Nanjing Traditional Chinese Medicine Hospital Affiliated of Nanjing University of Traditional Chinese Medicine,Jiangsu Nanjing 210001,China;Nanjing University of Traditional Chinese Medicine,Jiangsu Nanjing 210001,China)

机构地区：[1]南京中医药大学附属南京中医院,江苏南京210001 [2]南京中医药大学,江苏南京210001

出　　处：《中医药临床杂志》2024年第1期104-109,共6页Clinical Journal of Traditional Chinese Medicine

基　　金：国家自然科学基金(81904112);江苏省自然科学基金(BK20190136);南京市卫生科技发展专项医学重点科技发展项目(ZKX20050)。

摘　　要：目的:基于关联规则及无监督分类方法探寻孟河医派治疗中风病的用药规律,为临床中风病的中医药诊疗提供参考。方法:检索收集清初以来孟河医派丁甘仁、费伯雄、马培之、邓星伯、杨博良等9位代表性医家,归纳整理其诊治中风病的医话医案,采用关联规则及无监督分类方法对数据进行数据挖掘分析。发现诊疗过程中的四诊信息及与之对应的核心药物。结果:共纳入105首处方,涉及中药173味;首先通过简单关联分析,调整支持度及关联度获得前10位的药物,用药频次最高的为半夏,舌苔脉象出现频次最高的为苔腻。支持度最高的药对、舌脉分别为陈皮和半夏、滑脉和苔腻;使用无监督分类方法将中药及四诊信息自动分为4大类。结论:通过频数分析,简单关联规则及复杂关联规则以及无监督的分类方法逐层探寻出孟河医派对中风病的独特治疗特色,发现通过机器学习找出的四诊信息与药物之间有相互对应关系。结合舌脉分析与药物分析,孟河医家治疗中风以清热化痰、补益气血、平肝熄风、祛瘀化痰为主。Objective:To explore the medication rules of Menghe School of Medicine in the treatment of stroke based on association rules and unsupervised classification methods,and to provide a reference for the clinical diagnosis and treatment of stroke with traditional Chinese medicine.Methods:Search and collect 9 representative doctors from the Menghe Medical School since the early Qing Dynasty,including DING Ganren,FEI Boxiong,MA Peizhi,DENG Xingbo,and YANG Boliang,and summarize and organize their medical records on the diagnosis and treatment of stroke,using association rules and unsupervised Classification methods perform data mining analysis on data.Discover the four diagnostic information and corresponding core drugs in the diagnosis and treatment process.Results:A total of 105 prescriptions were included,involving 173 flavors of traditional Chinese medicine.First,through simple correlation analysis,the top 10 drugs were obtained by adjusting the support and correlation.The drug with the highest frequency was Pinellia ternata,and the drug with the highest frequency of tongue coating and pulse was greasy coating.The most supported drug pairs,tongue and pulse are tangerine peel and pinellia,Hua Mai and greasy coating respectively;the unsupervised classification method is used to automatically classify traditional Chinese medicine and four diagnostic information into 4 major categories.Conclusion:Through frequency analysis,simple association rules,complex association rules,and unsupervised classification methods,the unique treatment characteristics of stroke by Menghe Medical University were explored layer by layer.It was found that there was an interaction between the four diagnostic information found through machine learning and the drugs.Correspondence.Combining tongue and pulse analysis with drug analysis,Menghe doctors mainly treat stroke by clearing heat and reducing phlegm,replenishing Qi and blood,calming the liver and extinguishing wind,and removing blood stasis and reducing phlegm.

关 键 词：孟河医派 中风 关联规则 无监督分类方法 用药规律 

分 类 号：R255.2[医药卫生—中医内科学]