【作者】 张敏；

【导师】 罗军；

【作者基本信息】 重庆大学 ， 计算机科学与技术， 2018， 硕士

【摘要】 随着医疗信息化的发展,信息技术广泛的应用于医学领域的各个医疗环节。通过医疗信息系统管理医学过程在一定程度上提高了医疗服务的质量和效率。但是,这些医学信息却没有得到有效的利用,在这些信息上的操作主要是进行增删查改,实现信息记录的功能。为了使用户能便捷的获取医学信息,Web上出现了大量的医疗信息网站。这些网站主要提供疾病、症状、药品、病例等医学领域信息的关键词查询服务。这种查询服务使得普通用户能从网站中获取一些医学相关信息。但是基于关键字的查询服务的固有局限性导致查询服务经常出现无法查询、查询不准、查询不全等问题。病例作为疾病的治疗示例,蕴含了丰富的医学知识。用户希望能够快捷有效地获取与自己相关的病例,然后借鉴相关病例中的较为可靠的医学知识。基于关键字的病例查询服务无法满足用户的需求。本体是人工智能领域中的一个热点分支,广泛应用于知识工程、语义Web和语义检索等领域中。本体在语义和知识的层次上描述了概念与概念之间的关系,有利于计算机理解概念间蕴含的语义信息。本文的主要工作如下所示:(1)分析现有的Web医学资源和本体构建方法,提出了本体半自动构建方法,构建了较为丰富的医学领域本体。在研究现有的基于本体的语义相关度算法后,提出了基于本体的概念语义相关度算法。在此基础上提出了查询扩展方法,最后提出了文本相关度计算方法。(2)提出了基于病例相关度的病例推荐算法以及基于用户相关度病例推荐算法,为用户提供了一种新的医学知识获取方式。(3)设计实验,通过查准率、查全率及F值等指标评估病例个性化推荐算法。实验表明病例个性化推荐算法具有较好的性能,且具有较大的应用价值。更多还原

【Abstract】 With the development of medical informatization,information technology has been widely applied to various medical process.To a certain extent,managing medical processes through medical information systems improve the quality and efficiency of medical services.However,these medical information have not been effectively used.The operations on these information are mainly the functions of adding,deleting,checking and modifying to realize information recording.In order to enable users to easily obtain medical information,a large number of medical information websites have appeared on the Web.These websites mainly provide keyword search services for information on diseases,symptoms,drugs,medical records and so on.This inquiry service enables ordinary users to obtain some medical related information from the website.Nonetheless,due to the inherent limitations of keyword-based query services,it often appear to be inaccessible,inaccurate and incomplete.As a record carrier for medical staff’s medical activities,Medical record contains a wealth of medical knowledge.Users expect to quickly and effectively obtain medical records,and then learn more reliable medical knowledge in related medical records.Keyword-based medical records query service cannot meet the needs of users.Ontology is a hotspot branch in the field of artificial intelligence.It is widely used in the fields of knowledge engineering,semantic web and semantic retrieval.Ontology describe the relationship between concepts at the level of semantics and knowledge,which helps computers understand the semantic information contained in concepts.The main work of this paper is as follows:(1)Analyze the existing Web medical resources and ontology construction methods,propose a semi-automatic ontology construction method and build an abundant medical domain ontology.After researching the existing ontology-based semantic relevance algorithm,a concept semantic relevance algorithm and a text correlation calculation method are proposed.The experimental results are in line with expectations.(2)An ontology-based query expansion method was proposed to more accurately capture the users’ s query intent.(3)Disease case recommendation algorithms based on disease case relevance and user relevance were proposed to provide users with a new way of acquiring medical knowledge.(4)Designing experiments to evaluate the performance of personalized disease case recommendation algorithms by examining recall,accuracy and F-values.Experiments show that these algorithms have better performance and have certain application value.更多还原