National Basketball Association Player Salary Prediction Using Supervised Machine Learning Methods

Özbalta E., Yavuz M., Kaya T.

International Conference on Intelligent and Fuzzy Systems, INFUS 2021, İstanbul, Turkey, 24 - 26 August 2021, vol.308, pp.189-196 identifier

  • Publication Type: Conference Paper / Full Text
  • Volume: 308
  • Doi Number: 10.1007/978-3-030-85577-2_22
  • City: İstanbul
  • Country: Turkey
  • Page Numbers: pp.189-196
  • Keywords: Basketball, Machine learning, NBA, Random forest, Sports analytics, Supervised learning
  • Istanbul Technical University Affiliated: Yes


© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.Basketball is one of the most popular sports in the world and National Basketball Association (NBA) is the main figure for it. With the purpose of sustaining the balance between the basketball teams in the league, salary cap is implemented for all the teams in NBA. Considering the salary cap, decision makers of basketball teams should be careful while spending their budget. Since there are no transfer fees in NBA, salaries are the main expense for basketball teams. Therefore, determining the salaries of basketball players while making contracts is crucial to compose the best possible basketball team. In this research, dataset from the NBA 2K20 MyTeam video game and NBA players’ performance statistics of 2019–2020 season will be combined to predict the salaries for new contracts of NBA players by using machine learning methods. Shrinkage methods will be used to select best subsets. After, regression and decision tree models will be used to see which one produces the best mean squared error values. Results show that predicted salaries are very close to the new contract salaries of NBA players.