基于数据挖掘的郑钦安用药规律研究

Study on the medication regularities of ZHENG Qin-an based on data mining

目的利用现代统计学和数据挖掘方法,挖掘郑钦安的用药规律,为深入研究郑钦安的临床用药规律和学术思想提供统计学依据。方法从《医理真传》与《医法圆通》中收集组成明确的方剂。利用Excel建立方药数据库,采用SPSS 17.0和clementine 12.0软件对数据进行描述统计分析、双变量相关性分析、关联规则分析和聚类分析。结果方剂来源医著不少于3次者15种,来源于《伤寒论》的方剂最多,为42首。使用频次不少于10次的中药32味,排在前10味的中药主要是桂枝汤和四君子汤的药物组成。相关系数大于0.2的药对30对。关联规则分析提取出84条规则。聚类分析得出6组药和7对核心药对。结论发现郑钦安在用药上具有如下几个特点:(1)法宗伤寒,推崇仲景;(2)辨分二纲,药有阴阳;(3)善补脾胃,协和营卫;(4)配伍灵活,用药圆通。

Objective To excavate the medication regularities of ZHENG Qin-an by means of modern statistics and data mining methods, and provide a statistical basis for in-depth study of ZHENG’s clinical medication regularities and academic thoughts. Methods Prescriptions with specific composition were collected from the True Principles of Medical Science （ Yi Li Zhen Zhuan ） and Flexibility of Medical Method （ Yi Fa Yuan Tong） . Excel is used to set up the prescriptions database, SPSS17. 0 and clemen-tine12. 0 softwares were used to do descriptive statistical analysis, bivariate correlation analysis, association rules analysis, and cluster analysis. Results 15 prescriptions come from at least three medical books, prescriptions from Treatise on Cold Damages rank most, totalling up to 42. 32 kinds of Chinese medicinals are used not less than 10 times. The Chinese medicinals that rank top 10 are mainly the composition of Guizhi decoction and Sijunzi decoction. 30 herb pairs’ correlation coefficient is greater than 0. 2. Associa-tion rules analysis has extracted out 84 rules. 6 groups of medicinals and 7 core herb pairs have been ex-tracted by cluster analysis. Conclusion It is found that ZHENG’s drug application has the following char-acteristics：（1） his medication regularities follow Cold Damages, and respect Zhong Jing;（2） pattern di-agnosis and medicines both should be distinguished by yin and yang; （ 3 ） good at invigorating the spleen and stomach, coordinating nutrient qi and defense qi;（4） flexible compatibility and drug application.