数字人文视域下的青藏高原文旅知识图谱构建研究——以塔尔寺为例
Construction of Cultural and Tourism Knowledge Graph on the Qinghai-Tibet Plateau:A Case Study of Kumbum Monastery
李鑫豪 1赵维纳 1马龙龙 2安波3
作者信息
- 1. 青海师范大学计算机学院,青海西宁 810008;藏语智能信息处理及应用国家重点实验室,青海西宁 810008
- 2. 中国科学院软件研究所时空数据管理与数据科学研究中心,北京 100190
- 3. 中国社会科学院民族学与人类学研究所民族语言文化行为实验室,北京 100081
- 折叠
摘要
青藏地区素有"高原文化宝库"之称.然而受闭塞的交通条件和较滞后的经济水平的限制,青藏地区文旅资源的保护与弘扬工作始终处于滞后状态.为改善这一现状,该文以数字人文为导向,提出一套文旅知识图谱构建范式,在提示学习框架下采用联合学习的方式对文本中的实体与关系进行抽取,实现低资源条件下的知识抽取,并以全国重点文物保护单位"塔尔寺"为代表,完整地介绍了塔尔寺知识图谱从原始数据获取、本体设计、知识抽取到可视化展示的详细流程.最终,该文所构建的塔尔寺知识图谱共包含4 705个节点及17 386条关系.
Abstract
The Qinghai Tibet region is known as the"treasure trove of plateau culture".To improve the protection and promotion of cultural and tourism resources in the Qinghai Tibet region,this paper proposes a set of cultural tourism knowledge graph construction paradigm.Under the framework of prompt learning,it uses joint learning to extract entities and relations in the text,which realizes knowledge extraction under the low resource situation.Taking the national key cultural relics protection site Kumbum Monastery as an example,this paper introduces the knowledge graph of Kumbum Monastery from the whole process of original data acquisition,ontology design,knowledge extraction to visualization.The final knowledge graph of Kumbum Monastery constructed in this paper contains 4,705 nodes and 17,386 relations.
关键词
青藏文化/提示学习/联合抽取/塔尔寺知识图谱Key words
Qinghai-Tibet culture/prompt learning/joint extraction/Kumbum Monastery knowledge graph引用本文复制引用
出版年
2024
NETL
NSTL
万方数据