Betty Masruroh, Aji Prasetya Wibawa, Lisa Ramadhani Harianti, Christian Emeka Okafor
This study investigates the potential of unsupervised clustering algorithms to identify thematic and citation patterns in a national journal, with the aim of formulating effective development strategies. The dataset comprises 162 articles from Jurnal Pendidikan Geografi that have been cited by publications indexed in Scopus. Three clustering methods (KMeans, Agglomerative Clustering, and DBSCAN) were applied to article titles, publication years, and citation counts. Thematic mapping was conducted using TF-IDF vectorization, while citation-based clustering was performed to identify impact-related groupings. The results suggest the presence of clusters representing high-impact, low-impact, and thematically coherent groups. Among the methods, KMeans produced the most balanced and interpretable results, making it the most suitable for short-text bibliometric analysis. Based on these findings, the study offers practical recommendations for journal editors, including reinforcing prevailing themes, initiating targeted calls for papers, and refining promotional strategies in line with thematic trends. This research highlights the promise of lightweight machine learning techniques in advancing data-driven editorial planning, particularly for journals aiming to achieve international indexing standards. © 2025 IEEE.
Universitas Negeri Malang, Scientific Publication Unit, Malang, Indonesia; Universitas Negeri Malang, Electrical Engineering Department, Malang, Indonesia; Universitas Negeri Malang, Research Institute and Community Engagement, Malang, Indonesia; Nnamdi Azikiwe University, Department of Mechanical Engineering, Anambra State, Awka, Nigeria
Research at a Glance
Register to unlock
Topics & SDG Alignment
Register to unlock
Collaboration
Register to unlock
Author Profile (Selected)
Register to unlock
References Overview
Register to unlock
Journal & Source
Register to unlock
Metadata & Integrity
Register to unlock