[1Yin-B] Poster 2

[1Yin-B-57] Interdisciplinary knowledge structuring of global governance research using multi-objective optimization with BERTopicLarge-scale bibliometric analysis of scholarly literature

Yusuke Inoue¹, 〇Hina Ota² (1. Keio University, 2. Yamanashi Prefectural University)

Keywords:

Natural Language Processing,BERTopic,Multi-objective Optimization

This study aims to clarify the structural organization of contemporary global governance research through large-scale topic modeling. Using 13,205 academic articles published between 2020 and 2024 in the Scopus database, we apply BERTopic with a multi-objective optimization framework to ensure model robustness (coherence = 0.82; diversity = 0.87). The analysis identifies 74 interpretable topics spanning health, climate change, sustainability, and institutional reform. Co-occurrence network analysis reveals two principal thematic clusters—environmental and climate governance, and social system restructuring and resilience—while topics related to urban governance, climate adaptation, and the Sustainable Development Goals occupy high betweenness centrality positions, functioning as structural bridges. UMAP-based community detection further delineates coherent subfields across the literature. These findings provide an empirical mapping of the knowledge structure of global governance research and demonstrate its internally differentiated yet interconnected thematic configuration.