Knowledge Graph-Enhanced Course Recommendation System: A Multi-Relational Semantic Approach

Abstract

India's New Education Policy introduces unprecedented flexibility in higher education through multidisciplinary programs, multiple entry-exit points, and credit-based academic mobility. This paradigm shift necessitates intelligent systems to guide students through complex course selection decisions. This paper presents a novel Knowledge Graph (KG) enhanced course recommendation framework specifically designed for the New Education Policy compliance. We construct an Educational Knowledge Graph comprising 238 entities (122 courses, 110 skills, 6 career pathways) with multi-relational edges capturing course-skill associations, prerequisite dependencies, and career alignment mappings. Our recommendation algorithm leverages graph traversal techniques combined with semantic similarity measures to generate personalized course suggestions. Experimental evaluation on 28,785 real student trajectories from the Open University Learning Analytics Dataset (OULAD) demonstrates that our KG-enhanced approach achieves significant improvements: 30.2% over popularity-based baselines and 65.1% over random selection in NDCG@5. The study provides explainable recommendations, including skill-gap analysis and career-alignment justifications, supporting the vision of outcome-based, student-centric education. We release our implementation and processed datasets to facilitate reproducibility and further research in educational recommendation systems.