【作者】 曾令明；

【导师】 唐常杰；

【作者基本信息】 四川大学 ， 计算机应用， 2005， 硕士

【摘要】 中国传统医学经过几千年的发展,积累了数以十万首方剂。在这些方剂中隐含有大量的配伍规律,利用数据挖掘技术将它们发掘出来,对中医学的进一步发展具有重要意义。数据挖掘是从大量的数据中提取出隐含的知识、规律和行为模式的处理过程。药对药组挖掘是方剂多维数据分析(国家中医药管理局的课题)的基础,本文着重研究了中药药对药组的挖掘方法,主要做了以下工作: (1)用位图矩阵设计了方剂库的逻辑格式和存储模式,大大提高了对方剂库进行挖掘的速度。(2)针对中药配伍中“过分频繁反而变得平凡”的特性,提出了双支持度关联规则挖掘。(3)分析了基于支持度-置信度框架理论的关联规则挖掘方法存在的问题,并针对这些问题提出了双向关联规则挖掘。(4)开发了中药药对药组挖掘系统TCMCGMiner,并通过实验证明该系统能够快速全面地找到有趣药对药组。本文组织如下:第一章综述了数据挖掘的基本概念和有关技术。第二章介绍了数据挖掘在中药方剂研究中的现状和方剂多维数据分析的研究内容。第三章介绍了关联规则,分析了Apriori 算法。第四章描述了位图矩阵的建立及其操作。第五章阐述了双支持度关联规则挖掘及其算法。第六章阐述了双向关联规则挖掘及其算法,并进行了相关性分析。第七章介绍了中药药对药组挖掘系统TCMCGMiner 的结构及使用。第八章对TCMCGMiner 系统进行了速度测试和挖掘规则测试。第九章总结全文。更多还原

【Abstract】 Chinese Traditional Medicine (TCM) has developed for thousands of years and accumulated hundreds thousands of prescriptions which have lots of compatibility of medicines rules. Mining them using the data mining technology is very significant to the development of TCM. Data mining is a process that discovers knowledge from mass data and information. TCM couples and groups mining is the basis of the TCM prescriptions multidimensional analysis that is a project authorized by National TCM Administration. This work studies TCM couples and groups mining methods.The main contribution includes: (1) Designs the logical formats and physical storage model for TCM prescriptions database via bitmap matrix。(2) Proposes bi-support association rules mining. (3) Analyzes problems in the association rules mining based on support and confidence. Proposes bidirectional association rules mining. (4) Implements TCM couples and groups mining system named TCMCGMiner. Experiment results show that the TCMCGMiner can discover interesting TCM couples and groups fleetly and roundly. The paper is organized as follows. Chapter 1 summarizes data mining. Chapter 2 introduces the actuality of applying data mining in the research of TCM prescriptions and the content of the TCM prescriptions multidimensional analysis. Chapter 3 introduces the association rules mining, and analyzes the Apriori algorithm. Chapter 4 explains how to establish the bitmap matrix and how to operate it. Chapter 5 gives bi-support association rules mining and its algorithm. Chapter 6 proposes bidirectional association rules mining and its algorithm, and analyses the relativity of bidirectional association rules. Chapter 7 describes the framework and the use of TCMCGMiner. Chapter 8 gives experiments for mining speed and mining rules of TCMCGMiner. Chapter 9 draws a conclusion.更多还原