从叙词表到知识图谱的构建研究与实践

长久以来,图书情报领域致力于设计用于组织和管理大量知识信息的工具,这些工具通常被称为知识组织系统或者受控词汇表。由于多个社区和商业标准的长期独立开发,导致了各种不同的知识组织系统的出现,例如主题标引系统、叙词表、分类法等1]。叙词表具有用、代、属、分、参等特性,存储上具有网状关系,文章采用知识图谱结构对其进行存储,使其数据推理上具有较好的效果。文章主要对以1990年《机械工程叙词表—第二版》为基本对象2],将传统知识组织系统《机械工程叙词表》的数据重构、清洗成本文设计的SKOS结构的知识本体,使用数据存储技术,将《机械工程叙词表》的叙词转换为SKOS结构存储在图数据库Neo4j。这样的数据整理和处理,使得知识大型系统中的知识组织规范和管理大量信息数据成为可能。

Research and Practices From the Thesaurus to Knowledge Graph

The library and information science has long been dedicated to designing tools for organizing and managing large amounts of knowledge information. These tools are often referred to as "Knowledge Organization System"(KOS) or "controlled structured vocabularies". Due to the long-term independent development of multiple community and business standard, a variety of different knowledge organization systems have emerged, such as subject heading systems, thesaurus, taxonomies. The thesaurus has the relations like "U","UF","BT","NT"and "CT", which is organized as a network in storage. And This paper develops a method to store thesaurus in a knowledge graph structure, which also has a good effect in the data reasoning. In this paper, the main research was based on the 1990"Mechanical Engineering Thesauri"- Second Edition. By using data cleaning techniques, this paper reconstructed the traditional knowledge organization system "Mechanical Engineering Thesaurus" data into a SKOS knowledge ontology structure. By using data storage techniques, this paper stored the SKOS knowledge ontology structure in Graph database Neo4j. These knowledge organization systems make it possible to effectively organize and manage large amounts of information data.