【作者】 倪子建；

【导师】 荣莉莉；

【作者基本信息】 大连理工大学 ， 管理科学与工程， 2013， 博士

【摘要】 知识系统是具有高度复杂性和抽象性的系统。对知识系统建立模型是研究此类系统的有效方法。对系统建模的一般思路是研究组成系统的要素,以及要素之间的关系。按照关系所包含的域划分,可以分为二元关系和多元关系。当前研究中,能够反映知识系统二元关系特征的相关理论和方法被广泛的进行了研究,例如建模中的复杂网络模型,形式化语言中的OWL, RDF等。但是在实际知识系统建模中经常需要反映系统的多元关系特征,例如在建立科研合作模型中,需要描述某篇著作的研究者不仅只有两位；又如在知识的形式化表示中,需要描述“某件事情在某时某地发生”。这方面的相关研究显得相对不足。所以研究能够反映知识系统多元关系特征的相关理论方法在知识系统建模中是有意义的。本文在对相关研究进行综述的基础上,对知识系统建模中所涉及到的多元关系问题进行了分析,重点讨论了网络模型和语义模型在知识系统建模中可能遇到的多元关系问题。从研究整体思路来看,本文从理论研究入手,然后把相关理论方法应用于实际的知识系统建模中去。主要包括两个方面的内容。第一,重点研究在基于关系多边形的OWL中的多元关系表示问题,并把相关的理论方法应用到了应急预案语义模型中去。第二,从YAGO语言的多元关系表示问题入手,重点研究了面向维基百科内容的超网络模型构建的问题。具体的本文完成了如下工作：1.在说明了形式化语言的多元关系表示问题基础上,主要讨论了OWL的多元关系表示问题。不能直接表示多元关系是OWL的一个重要缺陷。在OWL中,具体化是表示多元关系的一般方法。具体化结构往往需要定义额外的概念和复杂的结构,容易造成错误。并不是所有多元关系都需要复杂的具体化结构。本章中,首先研究了一类特殊多元关系,即准多元关系的具体化方法。证明了准多元关系不需要额外定义概念和复杂的结构就可以进行具体化；进一步的,本文给出了识别准多元关系的方法,并且证明了准多元关系具体化结构的唯一性。然后把上面的相关理论扩展到一般的多元关系表示中去。在上述理论的基础上,最后讨论了检索问题,以及基于具体化的时间关系推理。2. Wiki作为一个巨大的知识库是知识共享的重要平台,从宏观角度研究Wiki中文章的知识属性和发展趋势,有利于了解科学技术和人文社会的发展方向。Wiki中文章的标签和概要信息栏中的信息相结合可以有效地概况这篇文章的主要知识,而Wiki内容本体可以形式化描述这些语义知识。我们通过研究发现Wiki内容本体有两个重要特点,存在多元关系和本体的规模不断增长。复杂网络模型无法建立上市模型。根据以上两个特点,在讨论了YAGO语言的多元关系表示问题的基础上,本文提出以超网络模型描述Wiki内容本体,进而研究这个超网络模型的演化模式。通过本章的研究探讨了多元关系在知识系统宏观属性和演化规律研究上的应用。3.为不同的应急主体,提供在不同状态下,应该进行的应急工作,是预案内容的重要组成部分。为了能利用计算机方便而快速得到上述类型信息,并且把众多的应急工作有序的组织起来,本文提出基于顶层本体的预案形式化模型。建立这个模型必须解决基于OWL多元关系表示,检索,时间关系推理等问题。依据上文所提出的相关理论,本章讨论了面向主体的应急预案语义模型。更多还原

【Abstract】 Knowledge system is a highly complex and abstract system. Constructing models for this kind of system is a effective way to study it. Generally, modeling system needs study elements and relation of it. If this relation links two elements, it is called binary relation. If more than two are linked, it is called n-ary relation. The theories which can reflect the binary relations properties of systems are widely talked in recently years. The example is complex networks, OWL (Web Ontology Language), RDF (Resource Describe Framework) and so on.However, the n-ary relation is one of important properties of knowledge systems. The ex-ample is that in research corporation network model the number of authors in one works may be more than two. Another example is that "one thing happened in sometime" needs ternary relations for representing. The studies in n-ary field are not enough. The purpose of this article is to fill this gap.After reviewing of relate works, the issues about n-ary in knowledge system model will be analyzed. The network models and semantic models willed discussed in this article. From the point of the research process, this article obtains from the theoretical research, and the related theory and method is applied to the actual knowledge system modeling. In this article, two problems are discussed. One is that the reification in OWL based on relation polygon is studies at first. The related theory is applied in the semantic model of emergency plans of China. The other is that the n-ary relation representing of YAGO is discussed. The hyper network of wiki article is constructed with the YAGO.The main contents of the paper are shown as follows.1. Reification is a universal method for describing n-ary relations in OWL. But the struc-tures of reified the n-ary relations need extra classes, predicates, individuals and axioms, which will make it quite unreadable and inconsistency in ontologies. The aim of our paper is to simplify the reification. A concept model called relation triangle is proposed firstly to find Apical Con-cept (AC for short) in ternary relations. Then, this method is extended to n-ary relations (n>3). The n-ary relation with AC can be reified by the combination of the n-1binary relation with-out defining new concept. And the results will not lose any information. Moreover, the Apical Concept Set can simplify reifications also. Our methods can reduce codifications in ontologies, save concept for reification and keep more semantic information in original statement. More importantly, our theories can guide domain experts in the reification of n-ary relations.2. Wikipedia as a largest knowledge base is an important platform for sharing knowledge in the world. It can help understand the developing trend in the field of science, technology and society through the macro-level statistical properties of the knowledge and their evolution regularities. The semantic information of the article in Wiki can be gotten from the labels and the infobox in the pages, which can be formally represented by ontology. Two important characters of Wiki ontology are proposed. One is there are n-ary relations in it. The other is the size of this ontology is enlarging. Moreover, we propose the evolution mechanism of the ontology of Wiki.3. An important part of Chinese disaster plans is to provide useful information for different emergency responders, which tells them what should be done in a certain kind of scenario. A formal model can help responders get these kinds of information from internet. In this chapter, we proposed a formal model based on upper ontology. The jobs are listed as follows:1) A domain ontology of Chinese plans is constructed.2) N-ary relations can be represented in our model.3) The users can retrieval the useful information from our model.4) Temporal reasoning can help to link the rescue works which scatter in different chapters and sections.更多还原