Personalized Learning Path Recommendations: Fusing Knowledge Graph Embedding, Sequence Mining, and Collaborative Filtering

The rise of online education, particularly through Massive Open Online Courses (MOOCs), has significantly broadened access to high-quality learning resources. However, these platforms still face ongoing challenges in maintaining learner engagement. To address this, researchers have developed advanced personalized Recommendation System (RS), suggesting courses and revealing relationships between learning resources. Given that learners engage with resources in diverse sequences based on their prior knowledge and goals, it is crucial to analyze individual learning behaviors and tailor recommendations accordingly. We propose a Personalized Learning Path (PLP) recommendation method that integrates Knowledge Graph Embedding (KGE) techniques, Collaborative Filtering (CF), and Sequential Pattern Mining (SPM) to recommend video resource sequences. Our approach constructs a MOOC-specific Knowledge Graph (KG), incorporating courses, video resources, and users to analyze learning paths. By vectorizing learners’ chronological patterns, our solution enables detailed representations of various learning trajectories and so can generate highly personalized recommendations. Through a series of experiments with real-world data from XuetangX, we demonstrate the effectiveness of our approach across several key metrics, outperforming state-of-the-art methods. The promising results confirm that our proposed method is highly effective in recommending PLPs and resources in MOOCs.