Predição do desempenho de jogadores da National Basketball Association

Abstract

The objective of this study is to explore an alternative for predicting the technical and tactical performance of athletes in the National Basketball Association (NBA). The focus will be on the individual performance of players, providing an alternative perspective to more common predictions, which are centered on team’s performance. Furthermore, existing and popularized individual performance indices will be presented and analyzed, allowing comparison with the Generalized Linear Model and Regression Tree models, which are adjusted with the objective of predicting the performances of players in new games. After the proposed models are fitted, each respective performance will be compared over different seasons, enabling more varied insights.