BERT-enhanced bibliometric mapping of scientific networks: insights from AI in education research

- Kuzminska, Olena (orcid.org/0000-0002-8849-9648), Smyrnova-Trybulska, Eugenia (orcid.org/0000-0003-1227-014X), Przybyła-Kasperek, Malgorzata (orcid.org/0000-0003-0616-9694), Smyczek, Filip (orcid.org/0009-0001-1363-0007) and Morze, Nataliia (orcid.org/0000-0003-3477-9254) (2025) BERT-enhanced bibliometric mapping of scientific networks: insights from AI in education research Information Technologies and Learning Tools, 6 (110). pp. 219-239. ISSN 2076-8184

Text Olena Kuzminska.pdf - Published Version Download (731kB)

Abstract

Understanding the structure of scientific knowledge and collaboration networks is essential for guiding research and policy decisions. Traditional bibliometric methods, based on keyword co-occurrence and network analysis, often face challenges related to semantic ambiguity and inconsistent terminology, which can obscure thematic patterns and collaborative links. This study examines the potential of BERT-based language models to address these limitations and improve bibliometric mapping. Using a dataset of 504 publications on artificial intelligence in education indexed in Scopus, we compared conventional VOSviewer clustering with BERT-enhanced preprocessing. The integration of BERT significantly improved semantic grouping by consolidating synonymous and morphologically varied terms, reducing keyword redundancy by 17% and increasing graph density, which resulted in clearer thematic clusters and more interpretable collaboration networks. These improvements revealed emerging research trends, such as ethical implications of AI and the growing role of generative models in education, while highlighting central institutions that act as global knowledge hubs. The findings demonstrate that BERT-based preprocessing not only enhances the accuracy and readability of bibliometric visualizations but also supports strategic decision-making in research management. Practical implications include the design of interdisciplinary curricula, the informed allocation of research funding, and the development of agile policies for the responsible adoption of AI. Beyond education, this approach can be applied to domains characterized by terminological complexity, such as healthcare, sustainability, and social sciences. By combining BERT with established bibliometric tools, researchers can achieve a cost-efficient, reproducible, and semantically robust method for mapping scientific landscapes and identifying collaboration opportunities.

Item Type:	Article
Keywords:	bibliographies; bibliometric analysis; clustering algorithms; BERT; data visualization; case study (AI in education).
Subjects:	Science and knowledge. Organization. Computer science. Information. Documentation. Librarianship. Institutions. Publications > 00 Prolegomena. Fundamentals of knowledge and culture. Propaedeutics > 004 Computer science and technology. Computing. Data processing > 004.9 Application-oriented computer-based techniques Science and knowledge. Organization. Computer science. Information. Documentation. Librarianship. Institutions. Publications > 02 Librarianship Science and knowledge. Organization. Computer science. Information. Documentation. Librarianship. Institutions. Publications > 3 Social Sciences > 37 Education > 37.01/.09 Special auxiliary table for theory, principles, methods and organization of education
Divisions:	Institute for Digitalisation of Education > Generic resouse
Depositing User:	Алла 1 Алла Почтарьова
Date Deposited:	27 Feb 2026 13:12
Last Modified:	27 Feb 2026 13:12
URI:	https://lib.iitta.gov.ua/id/eprint/748487

Downloads

Actions (login required)

View Item