Personalised learning path recommendation and knowledge tracing model for large-scale online education

This study proposes a unified framework that jointly models personalised learning path recommendation and knowledge tracing to improve individualised learning support in large-scale online education.The framework integrates learners' knowledge states, prerequisite relationships, learning load, and preferences within a single space, enabling dynamic tracking and coordinated optimisation.An online-updatable knowledge tracing model captures mastery levels, which inform a scoring and recommendation mechanism that adapts as knowledge states evolve.Experiments on the EdNet-KT1 dataset show the proposed model achieves superior prediction accuracy and lower mean absolute error than recent baselines, with reduced parameters and training time.This approach balances predictive performance and computational efficiency, offering a practical solution for personalised learning support.