历史人物本体构建及其查询推理研究

发布时间：2018-02-02 06:37

本文关键词： 历史人物领域本体构建规则库本体推理　出处：《广西师范大学》2017年硕士论文　论文类型：学位论文

【摘要】：信息检索、会话系统已经得到广泛研究和应用,例如常用的检索系统有百度、谷歌,会话系统小冰和Siri,以及服务行业的自动应答系统等等。随着智慧旅游热度的提升,人们越来越希望从检索系统或问答系统中获得更加详细和准确的旅游文化信息。国内在智慧旅游和历史文化信息检索、人物关系推理、信息查询方面有了一些研究,如《三国志》人物本体、旅游信息系统等,但是,历史人物关系推理和查询还需要深入研究。历史人物作为文化的传承者,其特殊的人物特性及超出常规的人物关系给人物的信息查询和推理都带来困难。历史人物本体构建的关键在于相关知识的获取和知识表示及推理。本文采用描述逻辑来形式化表示人物本体,刻画历史人物特征,定义人物的原子概念和原子角色,对人物关系定义语义关系,构建推理规则,提高语义准确度,进而实现隐含知识的推理。为了准确完备地构建历史人物本体和实现查询推理,本文首先提取历史人物在历史背景下的相关特征和社会关系,分类定义了人物的原子概念和原子角色,通过原子概念构造复杂概念,通过原子角色构建复杂规则,规范领域内的概念、属性和概念间的语义联系,从而构建历史人物本体模型,实现信息查询和推理。最后通过靖江王府实例的应用和研究验证了本文历史人物本体和推理机制的可行性,同时也为智慧旅游提供了智能推理机制。本文的主要工作包括以下几个方面:(1)研究基于描述逻辑的历史人物本体构建方法,构建符合历史人物特征的王府人物领域本体。DL具有较强的表达能力和可判定性,能够在语义层面上表达历史人物特征。定义原子概念和原子角色,连接DL中的构造算子来描述复合概念和复合关系。利用OWL编码和可视化构建工具Protege构建历史人物本体模型。(2)针对本体描述语言OWL DL推理能力不足的问题,本文提出基于SWRL(Semantic Web Rule Language)规则的查询推理算法,该算法运用SWRL规则刻画符合常识的历史人物关系规则,将历史人物概念间的逻辑关系转化成规则语言,通过本体结合规则的方式来提供更强大充分的知识表示能力和推理能力,增强对历史人物领域知识的演绎推理能力。(3)基于历史人物本体知识库和历史人物关系规则集,以靖江王府为应用场景设计实现了一个人物关系查询推理系统原型。使用Jena推理机解析本体中的上下位概念,转换本体模型中的人物相关知识,采用XSLT的方式转化规则格式。推理机整合本体知识库中的已知事实和自定义规则,推理得到隐含的语义查询结果,为SPARQL查询本体知识库提供丰富的本体语义联系。通过实验验证历史人物关系查询推理的可行性和有效性,实现靖江王府人物特征的可视化推理查询。
[Abstract]:Information retrieval, conversation systems have been widely studied and applied, such as the commonly used retrieval systems Baidu, Google, session system Xiaobin and Siri. As well as the service industry automatic response system and so on. Along with the wisdom traveling heat rise. More and more people want to obtain more detailed and accurate tourism cultural information from the retrieval system or question answering system. In China, intelligent tourism and historical and cultural information retrieval, persona relations reasoning. There have been some research on information query, such as "three Kingdoms" personage ontology, tourism information system, etc. However, historical personage relationship reasoning and query still need to be further studied. Historical figures as cultural inheritors. It is difficult to query and infer the information of the characters because of its special character characteristics and abnormal relationship. The key to the construction of historical personage ontology lies in the acquisition, representation and reasoning of relevant knowledge. This paper uses description. Logic is used to formalize the persona ontology. Describe historical characters, define atomic concepts and atomic roles, define semantic relationships, construct inference rules, and improve semantic accuracy. In order to construct the ontology of historical figures accurately and completely and to realize query reasoning, this paper firstly extracts the relevant features and social relations of historical figures in the historical background. It defines the atomic concept and the atomic role of the person, constructs the complex concept through the atomic concept, constructs the complex rule through the atomic role, regulates the concept, attribute and the semantic relation between the concept in the domain. In order to build a historical personage ontology model to achieve information query and reasoning. Finally, through the application and research of Jingjiang Wangfu case, the feasibility of historical personage ontology and reasoning mechanism is verified. At the same time, it also provides intelligent reasoning mechanism for intelligent tourism. The main work of this paper includes the following aspects: 1) study the ontology construction method of historical figures based on description logic. The construction of Wang Fu character domain ontology, which accords with historical characters, has strong expressive ability and determinability. It can express historical characters on the semantic level and define atomic concepts and atomic roles. Connecting the constructor in DL to describe the compound concept and compound relation. Using OWL coding and Visual Construction tool Protege to construct the Ontology Model of Historical characters. Aiming at the problem of insufficient reasoning ability of Ontology description language (OWL DL). This paper presents a query reasoning algorithm based on SWRL(Semantic Web Rule language rules. The algorithm uses SWRL rules to depict historical relationship rules which accord with common sense, and transforms the logical relations between historical figures into rule language. It provides more powerful knowledge representation and reasoning ability by combining ontology with rules. Enhance the deductive reasoning ability of historical personage domain knowledge.) based on historical personage ontology knowledge base and historical personage relation rule set. Based on the Jingjiang Wangfu, a prototype of the personal-relationship inquiry reasoning system is designed and implemented. The Jena inference engine is used to analyze the upper and lower concepts in the ontology and to transform the personal-related knowledge in the ontology model. The rule format is transformed by XSLT. The inference engine integrates the known facts and custom rules in the ontology knowledge base, and the implicit semantic query results are obtained by reasoning. It provides abundant ontology semantic connection for SPARQL query ontology knowledge base, and verifies the feasibility and validity of historical personal-relationship query reasoning through experiments, and realizes visual reasoning query of Jingjiang Wangfu character feature.
【学位授予单位】：广西师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.1;TP18

【参考文献】