G. Calvo, F. Palmí Perales, C. Armero, V. Gómez Rubio
Our study introduces a novel methodological framework for assessing basketball team performance using a longitudinal Bayesian network (BN) within the context of Bayesian hierarchical modeling. A BN is a probabilistic graphical model that represents stochastic relationships between variables as nodes and edges, encoding conditional dependencies. We developed this model by integrating key performance metrics such as shots made, attempted, fouls received, and minutes played, alongside player positions to assess team scores while accounting for individual variability among players. To illustrate our approach, we applied the model to the Philadelphia 76ers during the 2005-2006 season, analyzing data from 13 key players, including Allen Iverson and Andre Iguodala. Our findings demonstrate that this comprehensive framework offers valuable insights into team dynamics and scoring potential in basketball analytics.
Keywords: Bayesian Inference, Sports Analytics, Probabilistic Networks, Team Dynamics
Scheduled
Sports analytics
June 12, 2025 7:00 PM
MR 3