摘要
山水志是中国山水文化记忆的重要载体,结合数智技术探索中国山水志史料的知识重构与可视化路径有助于进一步激活其文化价值和记忆价值。文章围绕山水志史料,设计了以“物质维度—象征维度—功能维度”为主线的知识图谱构建方案,提出了涵盖资源采集、知识表示、知识抽取及知识可视化四个模块的知识重构与可视化研究路径。在实证方面,以《续修大岳太和山志》为例,基于大语言模型与低代码平台Huggingfists实现知识图谱全流程构建。通过可视化呈现武当山文化,实现其数字记忆场景重构。文章在山水志史料数字化开发与山水文化传承方面提供了有益的探索路径,也为历史人文领域引入大语言模型提供了可行方案。
As an important repository of Chinese landscape culture landscape chronicles serve as a significant cultural memory carrier Exploring the pathways for knowledge reconstruction and visualization with the aid of digital intelligence technologies can further enhance their cultural and mnemonic value This study focuses on landscape chronicles materials and designs a knowledge graph construction plan based on three semantic dimensions material symbolic and functional It proposed a research pathway for knowledge reconstruction and visualization covering four key modules resource collection knowledge representation knowledge extraction and knowledge visualization Incorporating the theory ofSites of Memory the study deepened the deconstruction and reconstruction of the cultural memory and natural landscapes embedded in the landscape chronicles This study used the Continuation of the Great Yue Mount Taihe Chronicle as a case study to conduct an empirical analysis The entire process of constructing a knowledge graph was implemented through the integration of a large language model and the low-code platform Huggingfists First knowledge extraction from landscape chronicle texts was performed through designed prompts Entities attributes and relationships within the text dataset were identified based on standardized conventions for knowledge representation Subsequently the extracted discrete knowledge was integrated and semantic associations between entities attributes and relationships were established through normalized descriptions of knowledge representation Finally the extracted and structured landscape cultural knowledge was stored in Neo4j as triples The dynamic knowledge graph built on the Neo4j graph database visually illustrated the interconnections between Wudang Mountains core landscapes associated historical events and cultural symbols This multi-perspective multi-level visualization enhanced the intuitiveness of landscape knowledge and provided comprehensive semantic context for cultural research Leveraging semantic analysis tools it supported the exploration of implicit knowledge and the study of cultural dissemination patterns In summary this study achieves a complete workflow from extracting and integrating data from landscape chronicle texts to constructing a semantic knowledge network through knowledge graph construction and visualization It transforms static flat knowledge in traditional historical materials into a threedimensional dynamic semantic network This approach offers valuable exploration pathways for the digital development of landscape chronicle materials and the preservation of landscape culture It also provides a feasible solution for integrating large language models into the fields of history and the humanities.
作者
周树斌
程双玲
ZHOU Shubin;CHENG Shuangling
出处
《大学图书馆学报》
北大核心
2025年第5期54-66,共13页
Journal of Academic Libraries
基金
内蒙古自治区社会科学基金项目“《中华民族共同体概论》文化遗产多模态知识组织与知识聚合研究”(编号:2025PTWY13)的研究成果之一。
关键词
山水志
知识图谱
大语言模型
知识重构
数字人文
Landscape Chronicle
Knowledge Graph
Large Language Model
Knowledge Reconstruction
Digital Humanities
作者简介
通讯作者:程双玲,邮箱:chengsl20@naver.com。
