Enhancing learner engagement through hybrid deep learning recommendation system in online education

In order to further assist the students in achieving?their career goals, a new hybrid recommendation system is presented using deep learning that combines advanced neural networks with content-based filtering (CBF) and collaborative filtering (CF) for course and resource recommendations. Leveraging the Open University Learning Analytics (OULAD) dataset, the model diligently studies multiple sources of data such as interaction?logs, performance numbers, and content metadata. Across several evaluations observed a substantial improvement of key interaction metrics and recommendation quality: AUC@5 improved from 0.41 to 0.58 (+78%), click-rate increased from 12 to 20 minutes (+67%), average session time?remained constant at 20min (but yielded an increase in the per-user average), while the latent number of facets grew significantly by +85% on average. Students?who followed the advice had earned 28% more credit-hours/units.” The excellent superiority of the hybrid model is shown by?its ability to increase interaction of users for e-learning and reduce dropout significantly (p