Personalized Blended Learning Resource Recommendations Using Graph Convolutional Neural Networks

作者：Mingyue Han · 发表于：Journal of Circuits Systems and Computers · 年份：2025 · DOI：10.1142/s0218126625503591 · 研究领域：Online Learning and Analytics、Recommender Systems and Techniques

As information technology rapidly progresses, blended online educational resources have gained widespread popularity in the educational sector, providing learners with unprecedented convenience and an abundance of materials. However, amidst this abundance of resources, the challenge of accurately recommending content that aligns with individual personalized learning needs has emerged as a critical issue requiring urgent attention. To tackle this issue, this paper presents a novel approach: a system for suggesting blended online educational resources utilizing Graph Convolutional Neural Network (GCNN).The essence of this approach lies in constructing a user-resource interaction graph to deeply explore the intricate relationships between users and resources. By leveraging the powerful capabilities of GCNN, it is possible to capture high-order relational information within the graph. This enables a more accurate understanding of users’ learning preferences and requirements, thereby facilitating precise evaluations of learning effectiveness and resource recommendations. The experimental findings reveal that, in contrast to traditional recommendation algorithms, this method demonstrates notable benefits in boosting user satisfaction and enhancing educational outcomes. The proposed GCNN-based approach not only provides learners with more personalized and precise recommendations but also contributes new ideas and methodologies to the intelligent development of the education field.