An Intelligent E-Learning Framework for Personalized Educational Recommendations in Secondary Education through Deep Q-Learning Optimization

Personalized learning has become an essential requirement in modern education,every student learns differently some need more time, some prefer visuals, and others engage better with interactive activities. Unfortunately, most traditional learning systems still follow a “one-size-fits-all” approach, which makes it hard to meet the unique needs of each learner. This research takes a step toward solving that problem by building an intelligent e-learning system that adapts to students using Deep Q-Learning (DQN), a modern reinforcement learning technique. In our approach, the student’s activities are treated as states, while the system responds with different teaching actions. A reward system is used to guide the agent so that it promotes learning progress, engagement, and motivation. We tested this model against two baselines: Tabular Q-Learning and a Rule-Based agent. The results were clear the DQN agent learned faster and performed better, stabilizing at rewards between 12 and 14, while Tabular Q-Learning leveled off at around 5–6, and the Rule-Based agent stayed fixed near 5.