Loading...

Predicting Football Player's Market Value and Determination Factors Affecting it Applying Machine Learning Methods

Fooladi, Mohammad Hossein | 2024

0 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 57268 (01)
  4. University: Sharif University of Technology
  5. Department: Industrial Engineering
  6. Advisor(s): Rafiee, Majid
  7. Abstract:
  8. In addition to being a popular sport in the world, football is a big industry and clubs are considered as economic enterprises of this industry. One of the most important decisions that the managers of these clubs must make is the buying and selling of players during the transfer windows, so that one decision can bring a profit or a loss of several million euros for that club.So obtaining a correct estimation of the market value and price of football players is considered a crucial requirement for club managers. With the advancement of science and the evolution of machine learning algorithms, researchers tried to create models to predict the market value of players by using the characteristics and statistics of players. In this research, using two data sets of FC 24 video game, players data set and clubs data set which includes 71 different features of 18250 players from all over the world, a model was created to predict the market value of football players and determine the factors that are important to predict it. the players are divided into 7 categories of goalkeeper, centre back defender, full back defender, central midfielder, attacking midfielder, winger and stricker based on their position and for each category linear regression, lasso regression, elastic network regression and decision tree are used to build a prediction model and the performance of these algorithms were compared with each other using several indicators such as R – squared and Mean Absolute Percentage Error. Finally, based on the elastic network regression algorithm that had the best results, a prediction model for the market value of the players was presented for each category. R – squared for this algorithm was 0.965 on average and Mean Absolute Percentage Error was 17.22%. The results of the modeling showed that the characteristics of the player's club are also important in determining their market value. Also, the division of players into 7 different categories improved the measurement indicators compared to previous researches
  9. Keywords:
  10. Football ; Machine Learning ; Player Market Value Prediction ; Football Transfer

 Digital Object List

 Bookmark

No TOC