Hybrid CNN-XAI Model for Predicting Student Academic Performance

This study presents a combined method that improves the ability to predict and understand outcomes by merging Convolutional Neural Networks (CNN) with Explainable Artificial Intelligence (XAI). The approach can pinpoint effective predictors by uncovering intricate patterns from behavioural data using convolutional neural networks (CNNs). XAI techniques ensure that instructors can gain valuable insights by clearly outlining the role of each component. The model greatly surpasses traditional methods, reaching an impressive 96% accuracy rate and a micro-average AUC score of 0.86 and achieved 95% precision, 95.5% F1-score, and a perfect Cohen's Kappa of 1.0, showcasing exceptional reliability. The results show how well the proposed system can generate interpretive data and its reliability in predicting grades. The hybrid CNN-XAI model empowers educators by enhancing their decision-making freedom, boosting prediction abilities, and aligning with contemporary teaching methods.