Λογότυπο αποθετηρίου

Πλοήγηση ανά Συγγραφέα "Pliatsikas, Anastasios"

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
Α Β Γ Δ Ε Ζ Η Θ Ι Κ Λ Μ Ν Ξ Ο Π Ρ Σ Τ Υ Φ Χ Ψ Ω
Φιλτράρετε τα αποτελέσματα πληκτρολογώντας τα πρώτα γράμματα
Τώρα δείχνει 1 - 1 από 1
  • Μικρογραφία εικόνας
    Τεκμήριο
    Predict shot success in NBA
    (2022-05-20) Pliatsikas, Anastasios; Πλιάτσικας, Αναστάσιος; Athens University of Economics and Business, Department of Management Science and Technology; Ntzoufras, Ioannis; Chatziantoniou, Damianos; Karlis, Dimitrios
    Basketball teams have expended significant (financial, human, and material) resources in an attempt to measure and identify court locations from which individual players shoot at a higher percentage. In this regard, the current study's main goals are to determine the best approach for predicting shot success as well as gain insights into the performance indicators affecting players' shooting effectiveness.Our sample included player tracking and play-by-play data from the NBA's 2014-2015 season, along with a variety of features (e.g. spatiotemporal, situational, player and team based attributes), allowing us to investigate a player's performance history and in-the-moment factors that may influence shot success. Six different classification models (decision trees, random forests, naïve Bayes, k-nearest neighbors, linear probability model and logistic regression) were developed and then evaluated in terms of accuracy and F1-score. This procedure was repeated once for all field goals, once for two-pointers, and once for three-pointers. We also investigated each team's shooting strategies, the factors influencing shooting efficiency and the defensive aspect of the game, while also focusing on LeBron James.Our results indicated that the optimal classifier for predicting the shot success was random forest, which was also the case for three-pointers. The logistic regression was the most accurate classifier for two-pointers and among the best overall. It was also highlighted that the spatial-temporal aspects of shots (primarily shooting distance) and the player's shooting abilities (mainly the average field goal percentage per shot zone and per shot type) had a significant influence on a shot’s success.Coaching staffs and general managers could use our models to better understand and improve their players' abilities and shot selection throughout a game, as well as to assist them in their coaching decisions, game plan development, creation of efficient lineups and acquisition of players who are complementary to the team’s needs.