基于关联规则挖掘的方剂配伍规律初步研究

Preliminary Study of Prescription Compatibility Law based on Association Rule Mining

目的:构建结构化的方剂信息数据库,挖掘方剂中药物之间的关联规则,进一步研究方剂的配伍规律。方法:运用基于正则表达式的信息抽取方法,以vs 2008编程实现自动抽取,并以SQL Sever 2005进行关联规则挖掘。结果:生成了结构化的方剂信息数据库,并初步抽取方剂信息81 305条;在最小支持度为10%,最小置信度为70%的条件下,得到高频药物9味,生成规则4条。结论:基于正则表达式的信息抽取方法使全面大规模研究方剂配伍规律成为可能,挖掘结果与中医理论和临床经验总体相符,为进一步研究提供也依据。

Objective： This paper aims to build a prescription information database, mining the rules of drugs in the prescription, and study the compatibility law. Method： Using information extraction based on regular expressions and vs 2008 programming to realize automatic extraction, and realized association rules mining in SQL Sever 2005. Result： Generated a structured prescription information database, and initial extracted prescription information 81 305. When the support degree is 10% and the confidence level is 70% , 9 kinds of high frequency drus and 4 rules were geted. Conclution： The method of information extraction based on regular expressions make it be possible that make a full-scale study of prescription compatibility law. The results of data mining and the theory of traditional Chinese medicine and clinical experience were matching overall, which lay a foundation for further study.