Predicting Player Performance in EA SPORTS FC 25: A Comparative Analysis of Linear Regression and Random Forest Regression Using In-Game Attributes

This study presents a comparative analysis of Linear Regression and Random Forest Regression models to predict player performance in EA SPORTS FC 25 using in-game attributes. The primary objective is to evaluate these models in terms of their accuracy and effectiveness in predicting player ratings based on key attributes like Ball Control, Dribbling, Defense, and Reactions. The dataset comprises 17,737 player records with multiple performance indicators, preprocessed to ensure quality data for modeling. The research process involves data exploration, model development, and evaluation using metrics such as Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and the Coefficient of Determination (R²). Results indicate that the Random Forest model outperforms the Linear Regression model, achieving a lower MAE and RMSE, and a higher R² score, highlighting its ability to capture complex, non-linear relationships among player attributes. The study’s findings underscore the significance of ensemble models in gaming analytics and provide insights for gamers and developers to optimize gameplay strategies and improve game mechanics. Limitations include data constraints, and recommendations for future research suggest incorporating more diverse player data and exploring advanced algorithms.