Student Performance Prediction Using Machine Learning

This study proposes a technically robust machine learning framework for predicting student academic performance, specifically targeting the final grade (G3), by leveraging the Portuguese student dataset. The approach incorporates comprehensive data preprocessing techniques, including categorical encoding, feature scaling, and dimensionality reduction via Principal Component Analysis (PCA), to enhance data quality and model performance. A Multi-Layer Perceptron Regressor (MLPRegressor) is employed to model the complex, non-linear relationships between academic, demographic, and behavioral input features and student outcomes. The system is deployed through a web-based interface developed using Flask, integrated with Gradio for interactive and real-time prediction capabilities. Additionally, an API layer is incorporated to support seamless integration with external platforms. This architecture supports data-driven educational interventions by enabling early identification of at-risk students and aiding institutions in implementing targeted academic support strategies. Index Terms: Student Performance Prediction, Machine Learning, MLP Regressor, Educational Data Mining, Academic Analytics