Exploring educational psychology mechanisms in course knowledge graph-based personalized learning support for vocational students

Introduction This study investigates the integration of educational psychology principles into a personalized learning support system based on course knowledge graphs, designed for vocational students. Traditional approaches to personalized learning often lack adaptability to individual cognitive and behavioral patterns, limiting their effectiveness in diverse educational contexts. Methods To address these limitations, the proposed methodology comprises three components: preliminaries, the Adaptive Knowledge Graph Learning Model (AKGLM), and the Personalized Learning Optimization Strategy (PLOS). The preliminaries formalize the problem and define the structural and semantic properties of the course knowledge graph. AKGLM dynamically adjusts to individual learning needs by employing graph based techniques to model student-specific requirements, while PLOS incorporates domain knowledge and insights from cognitive and behavioral theories to optimize learning paths and content delivery. Results and Discussion Experimental results on benchmark datasets demonstrate consistent improvements in recommendation performance and indicate the effectiveness of integrating graph based personalization with behavioral modeling. However, it should be noted that the adopted datasets do not explicitly include psychological factors, which are approximated using behavioral proxies in this study. The experimental findings primarily validate the contribution of knowledge graph and behavioral components, while only partially reflecting the impact of psychological modeling. The proposed framework highlights the potential of combining educational psychology with graph based methodologies to support personalized learning in vocational education. By aligning data driven approaches with human centered educational principles, this work provides a feasible direction toward more adaptive and scalable learning support systems, while acknowledging the need for more comprehensive datasets in future research.