Machine Learning Approach to Estimate Footbal...

Machine Learning Approach to Estimate Football Players Market Value

Author Name : Rohit Mittal

ABSTRACT

Thanks to the recent technological development in the field of Machine Learning, we were able to innovate and automate numerous applications, including healthcare, security, entertainment, and sports. In this research paper, I try to address the challenge of player valuation in Football. Football being the most competitive and revenue-generating sport, every team does a lot of crucial analytics before every decision. With the help of Machine Learning, I can attempt to predict player cost before the bidding event itself. This will help the teams to optimize their budget for the games. I will be applying various methods of Exploratory Data Analytics and Data Preprocessing followed by Regression Algorithms Such as LASSO, Ridge LightGBM, Deep Neural Network and XGB Regressor trying to find the optimal model for the real-world data (FIFA2021). After the detailed research and analysis, I identified that using log transformed data with XGBoost or LightGBM gives a better R2 score than classical approach like Polynomial Regression,

Keywords: XGBoost, LightGBM, LASSO Regression, Ridge Regression, Football, Market Value, Deep Neural Network.