A Learner Knowledge State Modeling Method Based on the Improved DKVMN
Yiqun Kou, Xi Zhang, Ye Yang, Youmin Hu, Bo Wu, Jiaxin Liu · 2023
In the field of modern industrial training, generic training methods often struggle to meet the personalized needs of various learners, leading to lower-quality training experiences and impacting the effectiveness of training outcomes. By incorporating customized learner parameters in the model, we can assess learners' knowledge mastery more accurately and develop more targeted training plans in subsequent training sessions. Therefore, this paper proposes a method for personalized knowledge state modeling of learners using an improved Dynamic Key-Value Memory Networks knowledge tracing algorithm(DKVMN-C). The comparison experiment results based on the AUC values of the DKVMN-C, the DKVMN, and the BKT indicate that this method effectively enhances the accuracy of knowledge tracing.