Deep learning-based personalized learning path recommendation system for higher education

The growing variety of student profiles in higher education has made it more important than ever to have smart systems that can change learning experiences on the job. Current recommendation algorithms fail to capture the complex relationships among learner context, behavior, and semantics. Therefore, this study, a Eduzyme-Net: Deep Learning and Enzyme Action Optimizer-Based Personalized Learning Path Framework, uses Graph Transformer Networks for relational learning, Federated Meta-Learning for adaptive cross-institutional generalization, and the Enzyme Action Optimizer (EAO) for evolutionary path refinement to overcome these constraints. Instead of using an attention-driven graph of learner-course interactions to model conceptual relationships and learning progress, the EAO network iteratively refines potential learning routes via affinity-based optimization and catalytic adaptation, inspired by enzyme–substrate dynamics in biology. Federated Meta-Learning protects personal data transferred across universities. A 23.4% increase in learning-path accuracy and 20.1% increase in learner satisfaction were reported in big academic dataset studies compared to standard deep recommendation systems. Results demonstrate that Eduzyme-Net can create context-aware, scalable, and customizable learning experiences. This enables autonomous, data-driven curriculum creation in higher education environments.