【作者】 佘莉；

【导师】 符红光；

【作者基本信息】 中国科学院研究生院（成都计算机应用研究所） ， 计算机软件与理论， 2006， 博士

【摘要】 学科知识是人类最基本的共性知识,如何利用现代信息技术进行有机地整合,为知识的学习者和传播者提供个性化服务,是知识工程和人工智能领域的重大问题,也是教育信息化的关键问题。Internet已经成为最丰富的知识资源之一,在教育领域也是如此。但在传统资源管理方式下,由于信息资源缺少统一的语义描述,用户难以查找到与需求相关的资源,难以实现相关资源的语义融合。如何使信息资源具有应用程序可以理解的含义,方便用户查找到与需求相关的信息资源,并根据信息资源所具有的领域知识,实现信息资源的语义融合,这些问题是信息资源管理所面临的挑战。知识本体是将网络环境下的大量非规范性知识进行概念化表示的一种有效手段,能够有效促进应用之间的概念语义共享。尽管知识本体技术已在诸多研究领域有着广泛的应用,如生物、医学、农业、制造工艺、数字图书馆等,但在学科知识领域中的研究还刚起步,本文尝试在这一新兴领域做出创新性探索。主要的工作和贡献有:1本文基于RDFS开发几何学科知识本体(Geometry Ontology),通过建立概念、概念之间的继承、实例以及多种约束关系表达统一标识的,可高度共享和重用的几何学科知识。2作为几何知识平台上的重要应用,分析和提出基于知识本体的几何概念相似性和知识相关性搜索流程和算法,并对相似度和相关度进行量化表示,这是进行智能推理和判断的重要基础。3针对平台上图形的作图特点和搜索要求,提出基于拓扑结构的相似几何图形搜索算法,具有对模糊性相似的图形识别的能力;并以编码的方式建立图形索引。4结合自然语言处理技术实现几何学科知识的自动获取,提出几何学科范围内的概念、断言和实例知识获取算法;实现几何知识组织的整体流程,包括文本映射、概念聚类和几何图形语义索引,提供自然语言驱动的搜索与查询。5实现学科知识平台架构及其上的基于Web的知识本体应用模式,提供诸如作图、搜索等知识服务。更多还原

【Abstract】 The discipline knowledge is the most fundamental common knowledge forhumanity. How to use modern information technology to integrate organicallyand provide personalized service for knowledge learners or disseminators, is amajor subject of knowledge engineering and artificial intelligence, also a key issueof information technology education.Internet has already become one of the most abundant knowledge resources,and the same to education field. But in traditional information resource manage-ment pattern, the information lacks a uniform semantic description. It is hardfor user to find more relevant information resource and realize the informationsemantic fusion.How to assign the managed resource a machine understandable meaning to fa-cilitate user to find what they really need , and according to the domain knowledgeto realize the information semantic fusion, are challenging tasks for informationresource management. The domain ontology is a formal, explicit specification ofa shared conceptualization, which can represent domain knowledge and facilitatethe inter-operation between the heterogenous application systems.Although the domain ontology technology has the widespread application inmany research areas, but the research in the disciplinary knowledge fields still juststarted. This article tried to make innovative exploration in this new developingfield. A significant part of the article concerns the following:1 Based on RDFS, the concepts in geometry domain and relations amongthese concepts, such as inheritance relationship, instance relationship andrestrictions, are built by the Geometry Ontology (GO) to represent highlysharable and reusable domain knowledge.2 As an important application of platform, the degree of concept similarityand relevant algorithm are proposed to solve the problems of computingsimilarity and relevant. It’s the important foundation of intelligent reasoningand judgment.3 To meet with the drawing characteristic of the figure and requirement ofsearching for on the platform, this article proposes the similar geometricfigure searching algorithm based on topological structure, which has theability to recognize fuzzy similar figures;and set up index by structurecoding.4 Combine with natural language processing technology to realize disciplinaryknowledge acquisition automatically, search and inquire in natural language.5 Realize the framework of disciplinary knowledge platform and applicationmodes based on Web for domain ontology, which o?er knowledge services asdrawing , searching etc..更多还原