基于数据挖掘探讨中医药治疗心脏神经官能症的组方用药规律  

Compatibility rule of traditional Chinese medicine for treatment of cardiac neurosis based on data mining

作　　者：刘祥 王超[2] 安莹 唐卓然[1] 齐鑫[1] 赵海滨[2] LIU Xiang;WANG Chao;AN Ying;TANG Zhuo-ran;QI Xin;ZHAO Hai-bin(Beijing University of Chinese Medicine,Beijing 100029,China;Dongfang Hospital,Beijing University of Chinese Medicine,Beijing 100071,China)

机构地区：[1]北京中医药大学,北京100029 [2]北京中医药大学东方医院,北京100071

出　　处：《海南医学院学报》2022年第16期1261-1267,共7页Journal of Hainan Medical University

基　　金：北京市自然科学基金资助项目(7202126);北京市中医药科技发展资金资助项目(JJ-2020-38)。

摘　　要：目的:运用“中医传承辅助平台(V2.5)”软件分析中医药治疗心脏神经官能症的组方及用药规律。方法:检索中国期刊全文数据库、万方数据库、维普中文期刊服务平台及中国生物医学文献数据库,依据纳入及排除标准筛选出运用中药方剂治疗心脏神经官能症的文献,利用中医传承辅助平台进行数据挖掘,统计用药频次及常用剂量、中药四气五味及归经频次。运用关联规则、复杂系统熵聚类分析法,提取核心药物组合,并采用无监督的熵层次聚类预测潜在新处方。结果:通过文献检索及筛选,最终录入210首方剂,通过“频次统计”共提取出181味中药,累计使用频次2 410次。其中,使用频次≥15次的中药共46味,根据功效主治,可大致包括理气药、安神药、补气药、清热药及补血药等。分析中药的四气五味及归经规律可知,温性、甘味药物最为常用,多归于心经,肝经和脾经次之。基于关联规则的组方规律发现,在药物组合中,白芍-柴胡应用最多;枳壳和柴胡的关联度最强,高达90%;网络图则提示,柴胡、当归、白芍、茯苓等药物为关联网络的核心。基于复杂系统熵聚类组合,可得到24个药物核心组合。运用无监督的熵层次聚类算法,提取出核心药物组合12种,进一步聚类得到6种潜在新方。结论:中医治疗心脏神经官能症以“气血同调”、“标本兼顾”、“补泻兼施”为原则,重视气郁的始动病机,主要从心肝脾入手,以疏肝理气、清热化痰、镇静安神、活血利水等为具体治法。Objective:To analyze the prescription and medication rule of traditional Chinese medicine(TCM)in treating cardiac neurosis by using the software of "Traditional Chinese Medicine Inheritance Assistant Platform(V2.5)". Methods:China Journal Full-Text Database,Wanfang database,Weipu Chinese Periodical Service Platform and Chinese biomedical literature database were retrieved according to the inclusion and exclusion criteria. Literatures on the treatment of cardiac neurosis with TCM prescriptions were screened out and data mining was carried out using the TCM inheritance assistant platform. Use frequency,commonly used dose,the frequency of properties,flavors and meridian distribution were counted. Association rules and entropy clustering analysis for complex system were used to extract core drug combinations and unsupervised entropy hierarchical clustering was used to predict potential new prescriptions. Results:Through literature retrieval and screening,210 types of prescriptions were finally selected,and 181 types of Chinese herbs were extracted,with cumulative frequency of 2 410 times. Among them,there were 46 traditional Chinese medicines that were used more than 15 times. According to the efficacy and indication,they were classified into qi-regulating drugs,tranquilizers,qi-tonifying drugs,heat-clearing drugs,blood-tonifying drugs,etc. According to properties,flavors,and meridian distribution,warm and sweet drugs were the most commonly used,mostly distributing to the heart meridian,followed by liver and spleen meridians. Association rules showed that combination of radix paeoniae and paeoniabupleurum was the most used;faructus aurantii and bupleurum had the strongest correlation,up to 90%. The network diagram indicated that bupleurum,angelica sinensis,radix paeoniae radix,and poria cocos were the core of the association network. Based on complex system entropy clustering,24 core drug combinations were obtained. The unsupervised entropy hierarchical clustering algorithm was used to extract 12 core drug combinatio

关 键 词：心脏神经官能症 中医药 中医传承辅助平台 组方用药规律 

分 类 号：R259[医药卫生—中西医结合]