您当前的位置:首页 > 论文详情

“国际智库如何叙事中国国家安全”:基于自建多源数据库的主题建模与多维图谱分析

“How Do International Think Tanks Narrative China's National Security”: Thematic Modeling and Multidimensional Mapping Analysis Based on Self-built Multi-source Database

摘要: [研究目的]基于多源异构数据融合技术、文献知识图谱动态更新机制设计自主构建国际智库研究数据库,为解构国际战略叙事的内在逻辑、预判复合型安全风险演化路径提供数据支撑,以增强新形势下国家安全治理的战略预判能力,为建设中国特色的优质智库数据库提供了独特的视角和资源。[研究方法]采用MySQL关系型数据库作为底层数据管理系统,结合自然语言处理技术与深度学习方法,实现数据的格式统一、图表统计和趋势预测。在此基础上,进一步筛选出2013-2023年间的1035份样本,融合文献计量和BERTopic主题建模方法,综合分析国际智库围绕中国国家安全的议题构成与认知结构。[研究结果]数据库分为源数据采集与存储、数据处理与标准化、分析与可视化、导出、安全与隐私五个部分,为科学知识图谱分析与深度学习主题建模提供了多维度研究支撑。而基于数据库的协同分析系统揭示了国际智库中国国家安全研究的叙事结构与演化脉络,不仅厘清了核心议题的聚类关系,还识别出潜在热点趋势,为战略研判与前瞻预警提供了实证支撑。

Abstract: [Research purpose] Leveraging multi-source heterogeneous data fusion and a dynamic literature mapping mechanism, this project aims to independently build an international think tank research database. It supports analysis of strategic narratives and prediction of evolving security risks, enhancing national security foresight and offering a distinctive, China-focused think tank resource. [Research method] A MySQL relational database underpins the system, integrating NLP and deep learning for data unification, charting, and trend prediction. Based on this, 1,135 samples from 2013–2023 are analyzed using bibliometrics and BERTopic to explore the topic landscape and cognitive structure of international think tanks on China’s national security. [Research result/conclusion] The database includes five modules: data collection, processing, analysis, export, and security. It supports multi-dimensional research through knowledge mapping and deep learning topic modeling. The collaborative analysis system reveals the narrative structure and evolution of China’s national security research in international think tanks, clarifies core topic clusters, and identifies emerging trends, offering empirical support for strategic insight and early warning.

版本历史

[V1] 2025-04-15 09:25:37 PSSXiv:202504.01932V1 下载全文
点击下载全文
在线阅读
许可声明
metrics指标
  •  点击量61
  •  下载量17
  • 评论量 0
评论
分享
邀请专家评阅
收藏