论文标题:知识的本体构建及检索模型研究
Ontology Construction and Knowledge Retrieval Model Research
论文作者
论文导师 马蕾,论文学位 硕士,论文专业 图书馆学
论文单位 四川大学,点击次数 121,论文页数 86页File Size4077K
2007-05-01论文网 http://www.lw23.com/lunwen_109606867/
Knowledge retrieval ;;Ontology ;; Knowledge base;; Knowledge representation ;; Concept knowledge;; Machining Process Technology
随着经济的发展，企业以及图书馆等机构的文献数量迅速增加。信息资源的增长伴随的是人们信息需求水平的变化，用户对信息质量的要求越来越高。这给情报检索技术带来了巨大挑战，因此新的理论和方法被引入到信息管理领域，其中之一就是本体论以及基于本体论的知识检索技术。 本体是概念化的明确的规范说明，从本质上，本体论就是某一领域中的概念以及这些概念间关系的集合，关系反映了概念间的约束和联系。 本体论已经成功地用于人工智能领域(AI)的知识表示。利用本体论理论和技术，把显性或隐性知识转变得可以被机器轻易识别，并使系统具有自动学习和推理能力，将大大提高用户检索信息的查全率和查准率，在解决信息检索中存在信息冗余、检索效率低下的问题上有着很大的优势，使得知识的管理更高效优质，其应用价值不言而喻。本文将讨论本体论的知识表示、构建理论以及基于本体的知识检索。 笔者在调研了国内外大量文献和网络资源(包括使用软件工具)的基础上，论述知识检索概念，与信息检索、语义检索的同异；阐述本体的起源、概念、类型，流行的本体知识库，以及基于本体的知识检索过程；详细介绍了本体论的知识表示理论，包括基于知识表示元语和方法，即利用本体元语进行知识的形式化描述。 为了更好地阐述本体知识表示的原理和方法，本论文选取了一个学科领域——机械加工工艺学作为入口，以该学科作为本体构建的领域知识，通过提取其核心概念集并建立概念关系，举例说明领域概念的获取以及建立概念间关系的方法，这是建立一个本体的关键步骤之一。同时，还详细说明了本体的构建理论和方法，包括一般性原则，主要的构建方法、本体语言和重要的本体编辑工具。 论文详细探讨了基于本体论的知识检索的基本原理、模型、检索策略及方法。对一个通用的基于本体论的知识检索模型进行探讨时，笔者就其中一些理论阐述了自己的不同认识，并作了一定的改造，进而完善模型。在这个基础上，论文介绍了一个国内的知识检索系统原型——Kretrieval。 目前，国外已经有了例如OntoFeb、OntoKnowledge这样较为成功的基于本体论的知识检索系统。相比之下，国内在本体论方面的研究刚刚起步，对本体论用于知识检索的相关方法和模型研究都十分缺乏，但经过国内科研研究人员的努力已经取得了一些进步，Kretrieval就是其中一项成果，这是国内的研究人员开发设计的一个基于本体的知识检索实验系统原型。Kretrieval系统是国家自然科学基金资助、同时也是教育部人文社会科学研究重大项目“信息可视化与知识检索(IVAKR)”的实验系统之一。笔者将就KR模型的知识获取、组织原理和实现方法，以及KR系统模型及其检索原理和过程进行详细说明，并对该系统进行评价。 基于本体论的知识检索系统具有很高的应用价值，将这种技术引入现代图书情报工作，将是未来的趋势。
The information and document in enterprise or libraries has increased explosively with the development of economy. The increasing number of information resources changes the level of people"s information demand in modern society. The developing level gives retrieval technology a great challenge, thus new ideas and methods are introduced into the field of information management to reform its theory and technology. So comes one of the hot topic-Ontology and ontology-based knowledge retrieval (KR), which is discussed in this paper. An ontology is an explicit specification of a conceptualization. In principle, ontologies are sets of concepts and relations which reflect context and binds of concepts. Ontologies perform well as an advanced technology of knowledge representation in Artificial Intelligence (AI). Through the ontological theory and technology, system can transform dominant or recessive knowledge easily distinguish by the computer, and performs to infer, greatly enhances accuracy ratio of retrieval information, make knowledge management more effective and higher quality. This paper discusses ontology building and ontology-based Knowledge retrieval model. The author intensively investigates and surveys large number of documents and web resources (ex. text paper and software tools).The theory and technology of ontology have been described, including origin, definition, type, popular ontologies knowledge base and knowledge retrieval procedure. Furthermore, the paper in detail introduces the elements and process of ontology-based knowledge representation, namely using the elements of representation to formalize knowledge. Specially, for demonstrate the procedure, taking a simple example of knowledge representation, the author selects the knowledge of machining processing technology as the field, and extracts the domain"s concepts, establishes the set and relations among them, thus constructs its ontology-based concepts set. At the same time, the theory of ontologies building has been introduced, including the general principle, several methods, ontological language and tools. The comparisons have made among 4 methods of building ontology, 12 ontology specification language, and 4 tools for editing ontology. The paper discusses the model, strategy and methods of ontology-based KR system. In some respect, the author explained the different views, makes some change and optimizes the model. On the basis, the paper introduces a domestic system prototype-Kretrieval. Nowadays, western researcher has developed some successful knowledge retrieval system, such as Onto-Feb and Onto-Knowledge. In comparison, study on this at home is just starting. The relevant knowledge about ontology for the retrieval is the lack of research, but our researcher is making much progress on this. "Kretrieval" is one success. "Kretrieval" is a National Natural Science Foundation-funded projects, and one of prototype experimental system of Education major in humanities and socialscience research projects-"IVAKR(Information Visualization and KnowledgeRetrieval )".The author explains "Kretrieval" knowledge acquisition, organization principle and implement method, then evaluates the model. With high application value, ontology-based KR system will be rapidly introduced into modern library and intelligence science as the tendency.
|
