Enhancing Algorithmic Fairness in Student Performance Prediction Through Unbiased and Equitable Machine Learning Models

The rapid adoption of ML algorithms has spurred the development of educational applications aimed at enhancing teaching and learning experiences. However, contemporary research underscores ethical concerns regarding their real-world implementation. A significant challenge lies in identifying and addressing potential biases in prediction models to mitigate their impact on diverse minority groups, including ethnicities, disabilities, nationalities, and genders. This paper conducts a thorough examination of algorithmic fairness, focusing on a detailed comparative analysis of traditional machine learning methods within a Student Performance Prediction (SPP) application for CS1 programming courses. The insights derived from this investigation not only enrich the ongoing discourse surrounding algorithmic bias and fairness but also pave the way for refining the development of just and equitable ML models.