Leveraging Neural Matrix Factorization (NeuralMF) and Graph Neural Networks (GNNs) for Enhanced Personalization in E-Learning Systems

This study investigates the application of a combined approach utilizing Neural Matrix Factorization (NeuralMF++) and Graph Neural Networks (GNNs) to enhance personalization in e-learning recommendation systems. The primary objective is to address significant challenges commonly encountered in recommendation systems, such as data sparsity and the cold start problem, where new users or items need prior interaction history. NeuralMF++ leverages neural networks in matrix factorization to capture complex non-linear interactions between users and content. GNNs model intricate relationships between users and items within a graph structure. Experimental results demonstrate a substantial improvement in recommendation accuracy, measured by metrics such as Hit Ratio (HR) and Normalized Discounted Cumulative Gain (NDCG). Additionally, the proposed model exhibits greater efficiency in training time than traditional methods, achieving this without compromising recommendation quality. User feedback from several universities involved in this research indicates high satisfaction with the recommendations provided, suggesting that the model effectively adapts recommendations to align with evolving user preferences. Thus, this study asserts that integrating NeuralMF++ and GNNs presents significant potential for broad application in e-learning platforms, offering substantial benefits in personalization and system efficiency