图书情报工作 ›› 2021, Vol. 65 ›› Issue (5): 126-135.DOI: 10.13266/j.issn.0252-3116.2021.05.013

• 知识组织 • 上一篇    下一篇

大规模中国历代存世典籍知识图谱构建研究

欧阳剑1,2, 梁珠芳3, 任树怀1   

  1. 1. 上海外国语大学图书馆 上海 201620;
    2. 上海外国语大学新闻传播学院 上海 201620;
    3. 广西民族大学管理学院 南宁 530006
  • 收稿日期:2020-08-12 修回日期:2020-12-21 出版日期:2021-03-05 发布日期:2021-04-14
  • 作者简介:欧阳剑(ORCID:0000-0001-5867-2852),研究馆员,博士,E-mail:oyjjj@163.com;梁珠芳(ORCID:0000-0003-3187-7502),硕士研究生;任树怀(ORCID:0000-0003-4817-407X),馆长,研究馆员,教授。
  • 基金资助:
    本文系国家社会科学基金项目"图书馆古籍文献的数字人文开发与应用模式研究"(项目编号:17XTQ003)研究成果之一。

Research on the Construction of Knowledge Graph of Large-scale Chinese Ancient Books

Ouyang Jian1,2, Liang Zhufang3, Ren Shuhuai1   

  1. 1 Shanghai International Studies University Library, Shanghai 201620;
    2 School of Journalism and Communication, Shanghai International Studies University, Shanghai 201620;
    3 School of Management, Guangxi University for Nationalities, Nanning 530006
  • Received:2020-08-12 Revised:2020-12-21 Online:2021-03-05 Published:2021-04-14

摘要: [目的/意义] 探索构建中国历代存世典籍知识图谱,以为研究者挖掘海量古籍书目数据背后隐藏的知识提供一站式平台,拓展古籍知识服务内涵,同时,大规模的典籍知识图谱也是机器智能的重要基础。[方法/过程] 通过知识图谱技术对中国历代存世典籍进行知识组织,从需求层、模型层、应用层3部分构建一个典籍知识图谱框架模型,通过人机协作进行典籍数据抽取及多源数据融合,完成数据的整理,并对典籍知识图谱实体类型及属性、典籍知识图谱实体关系及类型进行分析与定义。[结果/结论] 所构建的典籍知识图谱包含649 549种古籍实体、221 783位典籍责任者、1 498 383个古籍版本、13 960个地名节点,形成了一个立体、多维、多用途的古籍知识关联网络,对全球目前存世的主要中国历代典籍书目信息进行了较全面描述。

关键词: 古籍, 知识组织, 知识图谱, 人文研究, 数字人文

Abstract: [Purpose/significance] The establishment of a digital catalog is the need to protect and promote the Chinese civilization, and it also caters to the needs of new documentation and researchers. Chinese classics have been preserved throughout the ages. The construction of the knowledge graph provides a one-stop platform for researchers to dig out the hidden knowledge behind the massive bibliographic data of ancient books, which greatly enhances the knowledge service function of ancient books. The large-scale knowledge graph of ancient books is also an important foundation of machine intelligence. [Method/process] This research used knowledge graph technology to organize the knowledge of ancient Chinese classics, constructed a framework model of classics knowledge graph from three parts: demand layer, model layer, and application layer. Through man-machine collaboration, the data extraction of classics and multi-source data fusion, organize the data, analyze and define the entity types, attributes of the classic knowledge graph and the entity relationships, types of the classic knowledge graph. [Result/conclusion] It has realized the construction of the knowledge map of ancient books, including 649 549 kinds of ancient book entities, 221 783 persons in charge of ancient books, 1 498 383 versions of ancient books, 13 960 nodes of place names, and has formed a three-dimensional, multi-dimensional and multi-purpose knowledge association network of ancient books.

Key words: ancient books, knowledge organization, knowledge graph, humanities research, digital humanities

中图分类号: