基于领域本体与中文问句之间知识映射的研究

发布时间：2018-04-26 19:06

本文选题：领域本体 + 知识映射　；参考：《广西师范大学》2013年硕士论文

【摘要】：万维网的出现为人们提供了大量丰富的信息,并给人们的生活带来了许多的便利。而随着网络信息量爆炸性的增长,对于一个简单的咨询,人们已经被海量的网页索引所淹没而不知所措。因此人们急需要新一代的搜索引擎来诠释用户的问题并能在海量的网络信息中抽取客观的事实并通过筛选找出合适的答案反馈给用户。人们对知识的需求主要通过自然语言的方式来进行表达,而以问句为主体的问答系统是目前自然语言领域中研究的热点。它既能让用户以自然语言进行提问,又能为用户返回简洁、较为准确的答案。领域本体库的构建不仅可以清晰描述行业知识之间的概念及其关系,还可实现行业知识的共享和重用,有利于行业知识库的管理和维护,同时它对该领域中问句的解答有十分重要的作用。基于受限领域本体的中文问句的研究方法多数都是将人工构建的领域本体作为存储知识的知识库,针对问句进行语法分析、语义相似度的计算或者将它与句型模版进行匹配,再利用本体的查询和推理工具来反馈答案。但上述方法对问句自身语义的处理以及本体查询的精度仍然存在缺陷。为了克服上述不足,论文的第三章对领域本体的构建进行了更深刻的阐述,即对本体构建的局限性、人工构建本体、自动构建本体、半自动构建本体、以及本体的更新与维护进行的详细介绍,设计并实现了半自动构建本体的方法。论文的第四章对领域本体与中文问句之间知识的映射关系进行了深入研究,将中文问句答案的检索方法以及其对应方法下的问句类型进行了细分。针对问句答案的检索方法,作者将本体这种知识表示的方法与描述逻辑的背景知识联系起来,提出了基于TBOX的查询和推理方法、基于ABOX的查询和推理方法以及过程性知识查询的方法；而对于中文问句的处理,作者则提出语义框架与中词这两个概念,并很好的将它们融入到中文问句的语义分析中去。论文的第五章则是从工程实践的角度向读者进行受限领域专家系统展示,作者在第一现场分别对Jena的二次开发、SPARQL查询语言与Pellet推理机的应用以及测试问句集的获取进行了详细的叙述,并进行了相关的实验与结果分析。实验表明该专家系统在接受领域内的相关问题的提问时,具有较为合理的解释机制以及较高的回答正确率,能有效的应用于网络教学等其他领域。
[Abstract]:The emergence of World Wide Web provides a lot of information to people and brings lots of convenience to people ' s life . With the explosive growth of network information , people have been overwhelmed by massive page indexes . So people need a new generation of search engines to interpret the user ' s problems and extract objective facts in massive network information and to find suitable answers to feedback to users .

The demand of knowledge is mainly expressed in the way of natural language , and the question - and - answer system , which is subject to the question sentence , is the hot spot in the field of natural language . It can not only make the user ask questions in natural language , but also provide the user with concise and accurate answers . The construction of domain ontology library can not only clearly describe the concept and relationship among industry knowledge , but also realize the sharing and reuse of industry knowledge , which is beneficial to the management and maintenance of industry knowledge base , and it plays a very important role in answering questions in the field .

Most of the research methods of Chinese interrogative sentences based on restricted domain ontology are the knowledge base of storing knowledge , syntax analysis of questions , calculation of semantic similarity or matching it with sentence template , then using ontology query and reasoning tools to feedback the answer . But the above methods still have defects in the processing of the semantic of the question sentence and the precision of ontology query .

In order to overcome the deficiency mentioned above , the third chapter of this paper makes a more profound exposition of the construction of the domain ontology , that is , the limitation of the ontology construction , the manual construction of the ontology , the automatic construction of the ontology , the semi - automatic construction of the ontology , and the updating and maintenance of the ontology .
In chapter 5 , the author gives a detailed description of the application of Jena in the first field , the application of the query language and the inference engine and the acquisition of the test question set , and carries out the relevant experiments and results analysis .

The experimental results show that the expert system has a reasonable explanation mechanism and a high response rate when questioned by the expert system in the field of acceptance , and can be effectively applied to other fields such as network teaching and the like .

【学位授予单位】：广西师范大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP391.1

【参考文献】