【作者】 王姣；

【导师】 徐宝祥；

【作者基本信息】 吉林大学 ， 情报学， 2005， 硕士

【摘要】 本文首先回顾了哲学史上占有重要地位的本体论的演化历史,阐述了哲学范畴的本体论思想对于当代科学研究的重要意义。文中总结了信息科学领域对于本体的理解,在此基础上给出了本体的定义,分析了本体的组成要素、本体的分类以及特征,继而指出了本体在信息系统中的地位、作用以及应用。论文总结了本体构建方法论,介绍了已有的五种本体构建方法,并依据九种标准对这些方法进行了对比分析,找到了一点儿存在的问题和不足。继而引入软件开发中的RUP(Rational Unified Process,Rational统一过程)的概念,结合敏捷型方法的思想对RUP 进行了适当的裁减,对裁减后的RUP 命名为RRUP(Reduced RUP),在RRUP 的基础上探讨了一种本体建模方法,并给出了实施步骤。论文简要介绍了主要的本体描述语言,并且对其中的13 种核心的语言进行了五个方面的对比分析,提出了选择RUP 和RRUP 支持的UML(Unified Modeling Language,统一建模语言)对信息系统中的本体进行表达,并且尝试着给出了用扩展的UML 表达信息系统中的本体的几条指导性原则。最后,根据文中提出的基于RRUP 的本体建模过程以及与其相配套使用的扩展的UML,给出一个音乐本体的实例,简单说明了一下该过程实施和使用的方法。更多还原

【Abstract】 Ontology is an essential premise, philosophical viewpoint and theoretical base that any philosophy can’t avoid. As a basic theory, Ontology is applied in many specific sciences and really works in them. Ontology-driven information system development is the result that Ontology is used in information science and it chooses a method that uses the definitions as category to describe the real world, which is more abstract than the object-oriented technique. In this essay, the ontology modeling in the process of ontology-driven information development is the focus of this research. Ontology is not a new word but one of the ancient propositions in philosophy. Ontology researches are concerning about what exist and how they exist, that is to say, Ontology is a theory on existence and its nature. Positioned highly in philosophy history, Ontology is quite meaningful for the specific science research. Although in information science there is no general agreed definition for the ontoloty, in this essay, a well-accepted and extensively incited definition is still selected to describe what is ontology: Ontology is an explicit standardized formalized specification of a shared conceptualization. Based on this definition, the mathematical definition and model for ontology is put forward. Previously, people sort the ontologies according to their own experiences and criteria, but the understanding about the ontologies components can be still unified: ontology contains five components as category, slot, relation, axiom and instances respectively. Ontology plays a central role within an information system and affects the components in different aspects, such as application programs, information resources and user interfaces. Every component of an IS (information system) uses the ontology in its own way. The applications of ontologies for IS can be divided into four kinds such as neutral ontology, ontology as standardized specification of domain, ontology as a way to obtain information and ontology-based retrieval etc. There exist five kinds of methods to build an ontology as TOVE, Enterprise, Berneras et al., METHONTOLOGY and SENSUS. The essay contrasts the five methods according to 9 criteria and finds some problems such as work for unifying standardization for ontology modeling methods and process, lack of applicable guidelines for ontology modling and unsatisfied work about lifecycle process for the users. One idea this essay emphasizes to look on the ontologies as software products in the process of information system development, so RUP, which acts in the process of software development, can be used to build, model and represent an ontology. But while using the RUP to develop a system, especially to build an ontology, one must derive his own steps according his experience. Therefore, this paper discards the fussy approach before and adopts a method that can response rapidly to the changing of demand to reduce the process of RUP. At the same time the main charasteristics of RUP are remained and RUP after reduced is named RRUP. RRUP has cancelled the four steps of RUP and only given some workflows such as demand, analysis, design, implement and testing. The ontology modeling method based-on RRUP selects the interative life cycle and divides the ontology modeling process into three stages: specification, conception identification and formalization of the ontology model, ten steps. Ontology modeling and perfecting is a process that need superposition and evaluation time after time. That also accords with RRUP’s interative idea. Contrasting the main 13 description languages for ontologies in 5 aspects, this paper gets the following results: 1) there is none that suits the existing information systems modeling methods.2) visualization of some languages is not satisfied.3) every language has its own complicated methods and systems. Developers who are not familiar with these languages must try to learn them. Thus here UML, which is supported by RUP, has been chosen as the ontologies visualization to match RRUP. UML, as becoming a standard tool of systems modeling in software development, together with the semi-formalized OCL, not only is a feasible ontology modeling method but also has superiority to other knowledge engineering ones in: 1) developers are familliar with UML so更多还原