基于数据挖掘的中医无名方治疗带下病用药规律研究

Analysis on Medication Rules of Anonymous Prescriptions for Leukorrheal Diseases Based on Data Mining

目的:利用数据挖掘技术分析民间收集的中医无名方治疗带下病的组方用药规律,为发现新的临床有效方剂及新药发现提供新的途径,也为中医方剂形成过程研究提供实证基础。方法:收集整理治疗带下病的民间无名方,使用中医传承辅助平台(V2.5),采用频次挖掘、关联规则挖掘、基于无监督的熵层次聚类等数据挖掘方法,分析用药频次、常用药物组合、关联规则,挖掘潜在核心药物组合及候选新方。结果:收集并筛选出289首治疗带下病无名方,其中:使用频率较高的单味中药有:"鸡冠花"、"白果"、"山药"、"椿皮"、"黄柏"、"苍术"等;最常用的药物组合包括"山药-苍术"、"芡实-山药"、"黄柏-山药"等;关联规则分析则发现山药、芡实、黄柏、车前子及白果是最常使用的配伍,这与傅青主所立"易黄汤"组方相同。此外,尚生成7首新处方。结论:使用数据挖掘技术能从大量方剂中发现其组方用药的规律,且这些规律与中医名家的经典处方的用药规律相符合,这一方面可为新的临床有效方剂和创新药物发现提供基础,另一方面也证明中医经典方剂的形成是大量经验的积累,有其不可否认的临床实证基础。

Objective:To analyze the medication rules of anonymous prescriptions for leukorrheal diseases based on data mining, so as to provide a new way for discovering new clinical effective prescriptions and new drugs, as well asto provide an empirical basis for the study of formation process of TCM prescriptions. Methods:Anonymous prescriptions for leukorrheal diseaseswere collected and sorted out. Using data mining methods of Traditional Chinese Medicine Inheritance System(V2.5), such as frequency mining, association rules mining and unsupervised entropy hierarchical clustering, we analyzed drug use frequency, common drug combinations, and association rules, thenmined potential core drug combinations and new prescriptions. Results:A total of 289 anonymous prescriptions were collected. The most frequently used herbs in the prescriptions were Celosiae Cristatae Flos, Ginkgo Semen, Dioscoreae Rhizoma, AilanthiI Cortex, Phellodendri Chinensis Cortex, and Atractylodis Rhizoma. The most frequently used drug combinations were "Dioscoreae Rhizoma, Atractylodis Rhizoma", "Euryales Semen, Dioscoreae Rhizoma", and "Phellodendri Chinensis Cortex, Dioscoreae Rhizoma". The association rule analysis found that Dioscoreae Rhizoma, Euryales Semen, Phellodendri Chinensis Cortex, Plantaginis Semen and Ginkgo Semen were the most commonly used compatibility, this was the same prescription as "Yihuang Decoction" established by FU Qingzhu. And 7 new prescriptions were found. Conclusion:Using data mining technology, we could discover the rules of prescription medication from a large number of prescriptions, which were consistent with the classical prescriptions of famous TCM experts. It could provide a basis for the discovery of new clinical effective prescriptions and innovative drugs. On the other hand, it also could provide that the formation of classical Chinese medicine prescriptions was the accumulation of a great deal of experience and had undeniable clinical empirical basis.