Profiling temporal learning interests with time-aware transformers and knowledge graph for online course recommendation

Profiling users’ temporal learning interests is key to online course recommendation. Previous studies mainly profile by aggregating their historical behaviors with simple fusing strategies, which fails capture interest patterns underlying the sequential user behaviors. To fill gap, we devise a recommender that incorporates time-aware Transformers and knowledge graph better interests. First, introduce stacked extract enrollment sequences. In addition, design positional encoding module consider time intervals between courses. Third, incorporate utilize latent connections The proposed method outperforms state-of-the-art baselines for Furthermore, findings in ablation study offers several insights future research. model can be implemented platforms increase engagement reduce dropout rate.