Optimizing Personalized Recommender Systems for Teachers’ Digital Learning Models Using Deep Learning Algorithms

Abstract

Digital Learning Models for Teachers represent a significant advancement in the education sector, leveraging digital technology to enhance teaching effectiveness by creating personalized and customized learning experiences. These models enable teachers to receive more accurate feedback on their teaching. However, a current challenge is the need to recommend appropriate learning question types tailored to individual student needs, which can lead to suboptimal learning outcomes. To address this issue, this paper proposes a personalized recommendation algorithm based on Graph Neural Networks (PRAGNN) for teachers’ digital learning models. Specifically, the approach integrates DINA cognitive diagnosis and gray partial correlation evaluation to construct a student model that captures students’ mastery of knowledge points and cognitive ability levels. Additionally, a graph convolutional neural network (GCN) is employed, leveraging the sequential relationships between subject knowledge points to automatically capture semantic information from the higher-order structure of knowledge points, thereby enabling personalized recommendations. The algorithm achieved an accuracy of 85.7% through comparative experiments. This research presents a novel approach for developing personalized recommendation systems in teachers’ digital learning models, with the potential to significantly improve learning outcomes in the educational field.

Keywords

• Digital learning models for teachers
• deep learning
• graph neural networks
• personalized recommendations
• cognitive abilities