Fairness in Football Player Transfer Value Prediction

This project involves the development of a machine learning model designed to predict football players’ transfer values using a dataset that includes player names, ages, nationalities, positions, originating leagues, market values, and changes in market values (delta value 1 and delta value 2). The dataset will be complemented by actual transfer amounts for players who have been bought, allowing the model to be trained on real-world data. The primary aim is to examine potential biases and fairness issues within the model, especially focusing on variables such as nationality, age, and league. We plan to analyze how these biases manifest and propose strategies to mitigate them.