ORCID
Rui Yin, 0009-0001-0601-3676; Ningwei Sun, 0009-0003-4720-7498
Abstract
With the rapid advancement of artificial intelligence (AI) technologies such as ChatGPT, the use of knowledge graphs (KGs) has increased substantially in educational applications. As a technology for knowledge representation and management in AI, KGs play a pivotal role in personalized learning, educational resource integration and management, and learning performance prediction. Their scalability and semantic interpretability were particularly evident in online education during the coronavirus disease 2019 (COVID-19) pandemic. However, systematic reviews of their application in education remain limited.
To address this gap, our study analyzed data from the Web of Science (WOS) Core Collection (high-quality Social Sciences Citation Index and Science Citation Index Expanded journals) using a topic-based keyword search, which yielded 2,590 initial papers. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, the selection was refined to 117 high-quality articles focusing on empirical studies of KG construction and application in education. We employed three bibliometric tools: CiteSpace for analyzing temporal evolution and detecting emerging trends, VOSviewer for visualizing co-occurrence networks and conducting thematic clustering analysis, and the Bibliometrix R package for assessing scientific productivity performance. Analysis of the 117 articles identified core authors (Lanqin Zheng, Xiaona Xia, and Hao Xu), high-output countries/regions (China, United States, Australia), leading institutions (Beijing Normal University, Jilin University), and core journals in the field (IEEE Access, Applied Sciences–Basel, Education and Information Technologies). Keyword contribution analysis identified six research areas in KG-based educational applications. We found that KGs and education have undergone deep integration, with online learning emerging as a common application scenario. In particular, personalized learning path recommendation and resource recommendation are the primary application methods. Emerging technologies have provided new approaches for the integration of KGs into educational contexts.
Additionally, the study identified two emerging hotspots: predictive models and recommender systems. This study provides a clear framework for the application of KGs in education, contributing theoretical and practical insights for future research and implementation.
