基于知网的科学效应知识获取和本体库填充方法研究

发布时间：2018-12-13 17:37

【摘要】：产品创新是现代企业发展中必不可少的元素,其核心是功能的创新,本质是知识的创新。这个信息化高速发展的时代,人们对创新知识量的需求越来越大,那么首要解决的关键问题是寻找到能够快速且有效获取大量创新知识的方法和工具。科学效应知识是前人从大量的科学原理和专利中总结出来的,效应是发明问题解决理论(TRIZ)中一种基于知识的工具。该工具有助于创新设计领域中科学效应知识的表达,效应、功能之间的关系及其所在领域专业术语对创新思维起着重要的作用。科技文献具有创造性和实用性,已逐渐成为获取知识的焦点,但传统的信息抽取系统识别出来的信息有知识元、概念等,无法形成完整的知识体系,抽取的知识不能按一定的语义关系组织起来,形成完整的知识链存入知识库,在创新领域的应用性比较差。本文以科技期刊论文摘要为效应知识抽取对象,提出了基于知网语义相关度和基于句式模型匹配规则的科学效应知识抽取方法。分析科学效应知识本体库结构,归纳出效应知识的表达特征,同时大量分析了期刊论文摘要内容的知识表达特点,采用词法分析以及词性标注的方法,归纳出科技期刊文献摘要句式模型规则,最终从语义的层次上实现了科学效应知识的概念、功能、以及语义关系的抽取,并将所提取的知识按照三元组RDF本体的形式进行表达。本文构建了科学效应知识本体库,提出了将抽取到的效应知识与本体库中知识结构匹配的填充方法,从而实现效应知识本体库的自动填充。设计了一个科学效应知识抽取的系统原型,基本实现了期刊文档的相关知识抽取。通过对实验结果的分析表明,本方法具有一定的有效性和可用性。
[Abstract]:Product innovation is an essential element in the development of modern enterprises. Its core is the innovation of function and the essence is the innovation of knowledge. With the rapid development of information technology, people need more and more innovative knowledge, so the most important problem is to find the methods and tools to acquire a large amount of innovative knowledge quickly and effectively. Scientific effect knowledge is summarized from a large number of scientific principles and patents. Effect is a knowledge-based tool in the theory of problem solving (TRIZ). This tool is helpful to the expression, effect and function of the scientific effect knowledge in the field of innovative design, and the technical terms in the field play an important role in innovative thinking. Because of its creativity and practicability, scientific and technological documents have gradually become the focus of acquiring knowledge. However, the traditional information extraction system can not form a complete knowledge system because of its knowledge elements and concepts, etc. The extracted knowledge can not be organized according to a certain semantic relationship, forming a complete knowledge chain stored in the knowledge base, and its application in the field of innovation is relatively poor. In this paper, the abstract of scientific journals is taken as the object of knowledge extraction, and a method of knowledge extraction based on semantic relevance of knowledge net and matching rules of sentence model is proposed. This paper analyzes the structure of scientific effect knowledge ontology database, sums up the expression characteristics of effect knowledge, and analyzes the characteristics of knowledge expression of journal paper abstracts, and adopts the method of lexical analysis and part of speech tagging. The abstract sentence model rules of sci-tech journals are summarized. Finally, the concept, function and semantic relation of scientific effect knowledge are extracted from the semantic level, and the extracted knowledge is expressed in the form of triple RDF ontology. In this paper, the scientific effect knowledge ontology database is constructed, and the filling method is proposed to match the extracted effect knowledge with the knowledge structure in the ontology database, so as to realize the automatic filling of the effect knowledge ontology database. A prototype of scientific effect knowledge extraction system is designed, and the related knowledge extraction of journal documents is basically realized. The experimental results show that this method is effective and available.
【学位授予单位】：河北工业大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP391.1

【参考文献】