A Comprehensive Survey of Data-Driven Student Performance Prediction from Educational Data Mining to Explainable Artificial Intelligence

Abstract

Student performance prediction has become an important research area in digital education, focusing on analyzing academic, behavioral, and interaction data to improve learning outcomes and institutional decision-making. Artificial Intelligence, Educational Data Mining, Deep Learning, and Explainable AI techniques enable early identification of at-risk students while supporting personalized and adaptive learning environments. Existing studies apply machine learning models, neural networks, learning analytics, and explainable frameworks to achieve accurate prediction and effective academic support. This survey paper analyzes research contributions across major subtopics including Educational Data Mining, Machine Learning, Deep Learning, Peer Learning Analytics, AI-driven Feedback Systems, and Explainable AI by reviewing studies published between 2022 and 2026. The survey examines methodological trends, datasets, evaluation metrics, and performance outcomes reported in recent literature. However, challenges such as limited interpretability, scalability issues, real-time data integration limitations, and lack of cross-institution generalization still persist, highlighting the need for hybrid, transparent, and learner-centered predictive frameworks for intelligent educational systems.