A Course Recommendation System based on User Intents and Resource Alignment
Wenzhu Xie, Peng Pu · 2024
With the digital transformation of education in Chinese universities, a data-driven computer education revolution is emerging, aggregating high-quality courses and resources to support student development, while the phenomenon of ever-increasing AI-beyond-humans has triggered a new wave of discussion on how AI may change education. Students' demand for personalized course recommendation systems is increasing. However, existing recommendation systems face challenge of data sparsity due to limited user-course interactions, resulting in less than optimal recommendation performance. To effectively address this challenge, for the first time we introduce user intent in the field of personalized recommendations. Moreover, We have added knowledge graphs to enrich data information and proposed a resource alignment module based on the characteristics of course data. We propose a new model, KGIRA, based on these two modules. The effectiveness of our model was assessed using two course recommendation datasets, and a comparative analysis with current state-of-the-art models shows our model's superiority in several metrics, including NDCG. This indicates a significant advancement in the field of personalized course recommendation systems.