<![CDATA[This study aims to map the research landscape of Explainable Artificial Intelligence (XAI) for dropout prevention and learning loss mitigation. The study employed bibliometric science mapping to identify publication trends, conceptual structures, and emerging research directions. A Scopus search using the query TITLE-ABS-KEY("Explainable Artificial Intelligence" OR "Explainable AI") AND TITLE-ABS-KEY(Dropout OR "Learning Loss Mitigation") initially returned 207 records. No restrictions on language, document type, or publication year were applied during the retrieval process. Seven records with missing author metadata were excluded, leaving 200 documents published between 2020 and 2026 for analysis using Biblioshiny and VOSviewer. The results revealed rapid publication growth, with an annual growth rate of 57.04%, peaking in 2025 with 99 publications. Dominant themes included deep learning, explainable AI, learning systems, machine learning, dropout prediction, SHAP, LIME, interpretability, and uncertainty, while three broad conceptual clusters centered on deep learning, explainable artificial intelligence, and contrastive learning. Learning loss mitigation did not form a stable research cluster, indicating that the field has progressed more rapidly in explaining dropout prediction than in establishing connections between XAI and longitudinal learning decline. Theoretically, this study identifies a research agenda that integrates learning loss indicators with explainability, uncertainty, and fairness. Practically, it suggests that education policymakers should prioritize human-supervised early-warning systems whose explanations can be translated into timely and actionable interventions.]]>
Copyrights © 2026
