【作者】 樊凤杰；

【导师】 洪文学；

【作者基本信息】 燕山大学 ， 仪器科学与技术， 2017， 博士

【摘要】 方剂配伍知识发现是基于中医药理论,运用数据挖掘与知识发现等现代科技手段对传统中药组方配伍、药性配伍、组方结构等进行知识挖掘与发现的过程。将传统中医药理论与现代信息技术相结合,进行方剂配伍知识发现,对推动中医药理论发展,发挥中医药特色,促进中医药现代化具有重要意义。中医药数据具有整体性、复杂性及不确定性等特点,如何从大量的方剂数据中挖掘出符合中医药理论的共性与特异性知识是目前需要研究的课题。本文以中医药方剂配伍为背景,以形式概念分析及属性偏序为理论基础,研究基于数据属性偏序结构可视化的概念属性偏序结构方剂配伍知识发现方法,该方法将概念格与属性偏序结构图构建思想相结合,把概念格中的概念节点转变成属性偏序结构图中的支路,并从群结构、子群结构、支路和节点等对方剂配伍进行知识发现研究。首先,基于偏序理论及形式概念分析理论,给出概念格及属性偏序结构图生成原理及方法,研究形式背景净化、形式背景属性约简及形式背景分层优化方法,为数据偏序结构知识发现可视化研究奠定理论基础。其次,将概念格与属性偏序理论相结合,提出一种将概念格中的节点变成属性偏序结构图中的支路的知识发现可视化方法—概念属性偏序结构图,研究概念属性偏序结构图的生成方法。从概念属性偏序结构图的群结构、子群结构、节点及支路之间的关系等方面对中医药方剂配伍进行多层次、多角度、多层面的知识规则提取与知识发现。再次,以脾胃病为研究对象,构建《脾胃论》、《圣济总录》及《圣惠方》中方药形式背景,性味归经形式背景,并建立相应的属性偏序结构图,从属性偏序结构图节点层次、共有属性、伴生属性等分析医籍中治疗脾胃病的基础用药、高频用药及常用药组、药对,从群结构及子群结构分析方剂功效及性味归经,以揭示中医治疗脾胃病的遣方用药规律及药物组方模式。最后,以《伤寒论》方剂为数据源,挖掘《伤寒论》方剂组方结构配伍知识。基于属性偏序结构图研究七方中的大方、小方、奇方、偶方配伍规则提取方法,通过节点药物功效研究急方、缓方以及七情配伍知识发现方法,在计算药物相对计量基础上,进一步挖掘方剂君臣佐使配伍、君臣药物功效及药量等知识。更多还原

【Abstract】 The knowledge discovery on prescription compatibility is based on the theory of traditional Chinese medicine and pharmacy.Exerting the modern technological methods on data mining and knowledge discovery is the process of making knowledge mining and discovery to the prescription compatibility,property of a medicine and prescription structure.Combining the theory of traditional Chinese medicine and pharmacy with the the modern information technology to make knowledge discovery on prescription compatibility contributes to the development of the theory of traditional Chinese medicine and pharmacy,the characteristic of the traditional Chinese medicine and to the promotion of the modernization of the traditional Chinese medicine.Whereas,the data of traditional Chinese medicine and pharmacy is systematic,holistic,complex and uncertain,therefore,mining the knowledge with commonness and specificity from a large number of prescription data is the researching project of great necessity.This paper is based on the compatibility of Chinese medicine prescription and the theoretical basis of Formal Concept Analysis(FCA)and the attribute partial order.Knowledge discovery method on the prescription compatibility of concept attribute partial order structure based on visualization of the data attribute partial order structure is studied.This method combines the idea of the generation method of the concept lattice and attribute partial order structure diagram,and turns the node in concept lattice into the branch in the attribute partial order structure diagram.And the knowledge discovery of the prescription compatibility is completed by studying the relationship of the group structure,subgroup structure,node and branches in the concept attribute partial order structure diagram.First,the theory and the method of generating the concept lattice based on the partial order theory and formal concept analysis are introduced,and the purification method and the attribute reduction of the formal context in the perspective of redundancy of the attributes and objects are also given in this paper,which lays theoretical foundation for the knowledge discovery based on the data partial order structure.Besides,the concept attribute structure partial order diagram,a new visualization method for knowledge discovery and rule extraction,turning the nodes in concept lattice into the branches of the attribute partial order structure diagram,is proposed.The knowledge discovery and rule extraction of the prescription compatibility of traditional Chinese medicine and pharmacy are completed in a multilevel,multi-aspect way of various kinds by studying the relationship of the group structure,subgroup structure,node and branches in the concept attribute partial order structure diagram.Furthermore,the spleen and stomach disease are used as the studying object to compose the prescription-medicine formal context,the context on flavor and the meridian tropism and the attribute attribute structure partial order diagrams concerned based on Piweilun,Shengjizonglu,Shenghuifang are also generated.Then getting the basic medicine,high-frequency treatment,common medicine groups and pairs that are beneficial to treating the spleen and stomach disease by analyzing the node hierarchy,common attribute,associative attribute,the flavor and meridian tropism and the efficacy of prescriptions from the group structure and subgroup structure to discover the medication rules and medicine combination pattern.Finally,the prescription from Shanghanlun is used as initial data to mine the knowledge on compatibility of prescription structure of Shanghan Lun.The major preions,minor preions,odd-ingredient preions,and even-ingredient preions based on the attribute partial order structure diagram are studied.Besides,the knowledge discovery method on preions for the emergency,slow-acting recipe,seven relations is discovered by analyzing the efficiency of the medicine node.By computing the relative measurement of the medicine,the sovereign,minister,assistant and guide of the prescriptions,medicine and the dose of medicine are mined.更多还原