GKTP: A Personalized Knowledge Tracing and Predicting Method based on GNN
Feng Wang, Guangping Zhu, Xiaozhi Zhu · Procedia Computer Science · 2025
In the context of an information-driven society, the abundance and diversity of learning resources provide individuals with unprecedented learning opportunities, but they also bring challenges of information overload and cognitive overload for students. Particularly with the advancement of digital and intelligent education, the vast amount of learning resources often leaves students feeling overwhelmed and disoriented. Therefore, knowledge tracing and prediction of students’ learning processes are crucial for enhancing personalized learning settings and improving student performance. Traditional knowledge tracing methods have limited capacity in handling complex data and often lack adequate model interpretability. This paper proposes a knowledge tracing approach based on graph neural networks, utilizing knowledge graph modeling and tracing algorithms to monitor users’ knowledge states in real time. This enables real-time tracking and prediction of students’ learning processes, thereby meeting the needs of personalized learning.