图书馆杂志

图书馆杂志 ›› 2025, Vol. 44 ›› Issue (408): 70-83.

• 数字人文 • 上一篇    下一篇

星空记忆:中国古代天象记录智慧数据构建框架研究

唐振贵1 罗锦坤2 胡 蓉3 (1 广西财经学院新闻与文化传播学院 2 莆田学院新工科产业学院 3 西南大学教师教育学院)   

  • 出版日期:2025-04-15 发布日期:2025-04-24
  • 作者简介:唐振贵 广西财经学院新闻与文化传播学院,高级工程师,博士。研究方向:数字人文、语义网、知识图谱。作者贡献:总体设计、系统实现、论文撰写与定稿。E-mail:blackjack_edu@hotmail.com 广西南宁 530003 
    罗锦坤 莆田学院新工科产业学院,讲师,博士。研究方向:智能计算。作者贡献:数据收集与分析、算法设计。 福建莆田 351100 
    胡 蓉 西南大学教师教育学院,副教授,博士,硕士生导师。研究方向:信息检索与服务。作者贡献: 研究过程的指导与论文修改。 重庆 400715

Starry Sky Memory: A Framework for Constructing Smart Data from Ancient Chinese Celestial Phenomenon Records

Tang Zhengui1, Luo Jinkun2, Hu Rong3 (1 School of Journalism and Cultural Communication, Guangxi University of Finance and Economics; 2 New Engineering Industry College, Putian University; 3 College of Teacher Education, Southwest University)   

  • Online:2025-04-15 Published:2025-04-24
  • About author:Tang Zhengui1, Luo Jinkun2, Hu Rong3 (1 School of Journalism and Cultural Communication, Guangxi University of Finance and Economics; 2 New Engineering Industry College, Putian University; 3 College of Teacher Education, Southwest University)

摘要:

融合新一代AI 技术,探索中国古代天象记录智慧数据的构建框架,提升天象记录的数据形态,进而提供丰富的数智化应用,深挖其作为世界上独特的天文观测数据资源和珍贵的中国传统文化遗产的现代价值。以古籍资源智慧性保护体系框架为指导,融合知识图谱和大语言模型技术,构建涵盖史料层、注解层、关联层、融合层和应用层的古代天象记录智慧数据构建框架HDAA-SD。选择史料中的天象记录及其相关学术研究文献,使用HDAA-SD 框架构建古代天象记录的智慧数据,并通过对话模型来检验HDAA-SD 框架的应用效果。研究表明,HDAA-SD 框架有助于古代天象记录升级为智慧数据,并能助力古代天象记录在数智环境的应用开发,为古代科技文献的智慧数据构建和智慧化应用提供参考。

Abstract:

The study explores the integration of the new generation of AI technology to develop a smart data construction framework for ancient Chinese celestial phenomenon records. It aims to enhance the data format of these records, thereby enabling richer digital and intelligent applications while uncovering their modern value as a globally unique astronomical resource and a precious aspect of traditional Chinese cultural heritage. Guided by the framework for the intelligent preservation of ancient book resources, the study incorporates knowledge graph and large language model technology to construct the HDAA-SD (Historical Data Annotation and Application-Smart Data) framework. The framework covers five core layers: the historical data layer, annotation layer, association layer, fusion layer, and application layer. Celestial phenomenon records, along with related academic literature, are selected to test the construction and practical application of smart data using the HDAA-SD framework. The research shows that the HDAA-SD framework can help upgrade ancient celestial phenomenon records to smart data, supporting their integration into intelligent digital environments. The framework also provides a reference for the intelligent digitization and application of ancient scientific and technological documents.