中医舌象脉象本体构建及其应用示范研究

发布时间：2018-02-11 15:43

本文关键词： 舌象脉象本体构建规范化信息表达　出处：《湖北中医药大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着我国中医医院信息化建设的不断推进,医院信息系统采集了大量的医疗数据。大数据技术出现后,人们也更加重视这些来自日常医疗行为产生的数据,利用大数据技术对其进行处理、分析与利用,开展了大量的中医药临床相关研究,这也充分体现了中医从临床来到临床去的思想。舌象、脉象在中医诊断中发挥着重要作用,但受自然语言灵活性的影响,在描述上存在多种形式,计算机在识别、处理上会出现问题,影响了数据挖掘结果的质量。本体能形式化、规范化描述相关知识,在规范医学用语上已得到运用。因此,构建中医舌象、脉象本体,使其在规范舌象、脉象信息表达、提高挖掘结果质量与挖掘效率等方面发挥作用,是一项亟待开展的工作。目的:构建中医舌象、脉象本体,规范舌象、脉象诊断信息表达,为计算机识别、处理舌象、脉象信息奠定基础,规范和辅助医生进行舌象、脉象信息的采集。方法:按照七步法构建中医舌象、脉象本体。从已发布的标准、《中医药学名词》、《中医症状学研究》、《中医症状鉴别诊断学第二版》中收集舌象、脉象术语,进行去重和同义词整合处理。然后采用自上而下的方法初步确定分类框架,再从术语中提取属性值自下而上聚类,确定本体类、属性及约束,利用Protégé进行舌象、脉象本体的编辑。利用信息分类与编码技术,以属性组合的代码形式对舌象、脉象本体实例进行表示,并构建实例库。抽取四家三甲中医医院近几年的中医电子病历有关数据进行本体完整性验证,基于舌象、脉象本体开发中医舌象脉象规范化辅助软件。结果:根据分类框架和聚类结果划分了舌象、脉象本体类,舌象本体类属性有14个,脉象本体类属性有5个,舌象与舌象之间、脉象与脉象之间存在同义关系、上下位关系。舌象与脉象均采用定长主码与不定长补充码形式进行编码,舌象编码涉及本体类及其属性等13个方面,脉象编码涉及本体类及其属性等3个方面,构建了舌象、脉象本体实例库。对从中医电子病历中抽取的舌象、脉象诊断信息进行了编码验证,得到舌象总验证成功率为99.19%,脉象总验证成功率为99.52%。开发的系统能为用户提供规范的舌象、脉象诊断信息及主病的查询,规范和辅助用户进行舌象、脉象的采集,并能对舌象、脉象诊断信息进行编码,能导出相关数据,可用于数据挖掘中舌象、脉象数据的预处理。结论:构建的舌象、脉象本体能规范化表达舌象、脉象诊断信息,其具有完整性,能满足中医临床实际需求,促进了舌象、脉象信息的数据化、结构化。此外,本研究在促进标准应用、提高中医电子病历诊断信息质量方面进行了有益探索。但舌象、脉象的属性体系还需进一步完善,软件规范舌象脉象部分还需做成输入法的形式,以方便用户随时调用。
[Abstract]:With the continuous development of Chinese medicine hospital information construction, hospital information system has collected a large amount of medical data. After big data technology appeared, people also pay more attention to these data from daily medical behavior. Using big data technology to deal with, analyze and use it, and carry out a large number of TCM clinical related research, which fully reflects the Chinese medicine from the clinical to clinical thinking. Tongue, pulse in the diagnosis of TCM plays an important role. However, due to the flexibility of natural language, there are many forms of description. The computer will have problems in identification and processing, which will affect the quality of data mining results. Therefore, the construction of tongue image and pulse ontology of traditional Chinese medicine has played an important role in standardizing tongue image and pulse information expression, improving the quality of mining results and mining efficiency, etc. Objective: to construct the tongue image, pulse body, standard tongue picture, pulse diagnosis information expression of traditional Chinese medicine, lay the foundation for computer recognition, processing tongue picture, pulse image information, standardize and assist doctors to carry out tongue picture. Collection of pulse information. Methods: according to the seven-step method, the tongue picture and pulse image of traditional Chinese medicine were constructed. The terms of tongue and pulse were collected from published standards, "traditional Chinese Medicine nouns", "Research on symptoms of traditional Chinese Medicine", "second edition of TCM symptom differentiation Diagnostics". Then the classification framework is preliminarily determined by the top-down method, and then the attribute value is extracted from the term from the bottom to the top, and the ontology class, attributes and constraints are determined, and Prot 茅 g 茅 is used to carry out the tongue image. The editing of pulse ontology. Using the technology of information classification and coding, the tongue image and pulse ontology instance are represented in the form of attribute combination code. And build a case library. Extract the relevant data from four three A TCM hospitals in recent years for ontological integrity verification, based on tongue image. Results: according to the classification framework and clustering results, the tongue image, pulse ontology class, tongue image ontology attribute has 14, pulse image ontology attribute has 5, tongue image and tongue image. The relationship between pulse and pulse is synonymous, and the relationship between upper and lower. Both tongue and pulse are coded in the form of constant length master code and constant length supplementary code. Tongue image coding involves 13 aspects, such as ontology class and its attributes. Pulse coding involves ontology class and its attributes. The tongue image and pulse ontology case library are constructed. The tongue image and pulse diagnosis information extracted from electronic medical records of traditional Chinese medicine are coded and verified. The total success rate of tongue image verification is 99.19 and that of pulse image verification is 99.52.The system developed can provide users with standardized tongue picture, pulse diagnosis information and inquiry of main diseases, standardize and assist users to collect tongue image, pulse image, and tongue image. It can be used to preprocess the tongue image and pulse data in data mining. Conclusion: the constructed tongue image and pulse ontology can express the tongue image and the pulse diagnosis information. It can meet the actual needs of TCM clinical practice, and promote the digitalization and structuralization of tongue and pulse information. In addition, this study has made a beneficial exploration in promoting the application of standards and improving the quality of diagnostic information of electronic medical records of TCM. The attribute system of pulse needs to be further improved, and the software specification of tongue image needs to be made into the form of input method, so as to make it convenient for users to call at any time.
【学位授予单位】：湖北中医药大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13;R241

【参考文献】