Heterogeneous Graph Based Knowledge Tracing
Yingtao Luo, Bing Xiao, Hua Jiang, Junliang Ma · 2022
Recent advances in on-line tutoring systems have brought on an increase in the research of Knowledge Tracing, which predicts the student's performance on coursework exercises over time. Previous researches, such as Bayesian Knowledge Tracing, Deep Knowledge Tracing (DKT) and qDKT, focused on either skill-level or question-level. As a result, those methods fail to take question-skill correlations into account. Inspired by Heterogeneous Graph Embedding (HGE), We propose a HGE-based knowledge tracing model. In this paper, a heterogeneous graph is built on skill information and question information, so as to capture the latent interactions between skill nodes and question nodes. In the proposed method, the knowledge tracing model can leverage more informations than previous methods. The experimental results show that the proposed method outperforms other state-of-the-art methods centered on either skills or questions.