基于互信息的失眠用药关联性分析

基于中医传承辅助系统软件,分析郭蓉娟教授治疗失眠的用药经验。收集郭蓉娟教授治疗失眠的中药方剂,采用关联规则apriori算法、改进的互信息算法、复杂系统熵聚类等无监督数据挖掘方法,确定处方中各种药物的使用频次及药物之间的关联规则,分析郭蓉娟教授治疗失眠的用药经验。对筛选出的3 084首处方进行分析,确定处方中药物的使用频次,药物之问的关联规则,挖掘出41个核心组合和7首新处方。结论符合郭教授对失眠病因病机的认识:以肝郁化火为核心病机,肝郁脾虚、久病及肾的特点,中医传承辅助系统对于挖掘名中医临床经验具有重要实用价值。

Mutual information-based correlation analysis of herbs against insomnia

This paper aims to analyze Professor Guo Rongjuan's medication experience on insomnia therapy based on the Traditional Chinese Medicine (TCM) Inheritance Support Plat. First, TCM formulae prescribed by Professor Guo for insomnia therapy were collected from the TCM Inheritance Support Plat. Next, unsupervised data mining algorithms, including apriori, modified mutual-information, and entropy clustering of complex system were applied to obtain the frequencies for different herbs and identify the association rules among the herbs. Accordingly, we can gain new insights into Professor Guo's medication experience on insomnia therapy. Based on analysis of 3 084 formulae, we determined the frequencies for herbs in the formulae and identified the association rules among these herbs. At last, 41 core combinations and 7 new formulae were obtained.The identified medication experience conform with Professor Guo's views on the etiology and pathogenesis of insomnia: "pathogenic fire derived from stagnation of liver-QI (Gan Yu Hua Huo)" is the core pathogenesis of insomnia; "liver stagnation and spleen deficiency" and "chronic illness transferred to kidney" are the main features for insomnia. The TCM Inheritance Support Plat is of great practical value for mining clinical experience of famous TCM doctors.