PERSONALIZED LEARNING PATH GENERATION USING KNOWLEDGE GRAPHS AND NEURAL RECOMMEN SYSTEMS

This dissertation presents a hybrid model for personalized learning path generation using knowledge graphs and neural recommendation systems. The proposed approach integrates semantic reasoning from knowledge graphs with the predictive power of deep neural networks—specifically Graph Neural Networks (GNNs) and Transformer architectures—to recommend adaptive, individualized learning sequences. By modeling both conceptual relationships and learner behavior, the system dynamically adjusts to users’ evolving knowledge levels. Experimental results on educational datasets such as EdNet and ASSISTments demonstrate significant improvements in accuracy and personalization compared to traditional methods. The model also enhances explainability through attention mechanisms and graph-based reasoning. This research contributes to the development of AI-driven adaptive education systems, offering a scalable and interpretable framework for next-generation intelligent tutoring platforms.