Abstract:
Salary estimates for cricket players have always been made based on subjective judgment and scant evidence, leading to inconsistent and unequal pay. This study suggests a data-driven method for predicting player salary based on a variety of contextual and performance
elements using the XGBoost Regressor. Through a user-friendly interface made with Flask and React, the system provides salary estimates based on past data like batting, bowling, fielding, match conditions, and opponent strength. Users can choose players, see projected contracts, and assign them to appropriate contract bands. Explainability was measured using SHAP values, while model performance was monitored using traditional regression metrics. The solution demonstrates how using machine learning to determine cricket salaries can increase fairness and transparency.
