PRATA NETTO, L. G.; http://lattes.cnpq.br/3239756724529921; PRATA NETTO, Lourival Gonçalves.
Resumo:
This article presents a methodology for predictive analysis of sports game results, using National
Basketball Association (NBA) games as a case study. Specifically, the NBA is renowned for using data
analytics and advanced statistics to improve team and player performance. The objectives of this
work are to analyze the availability and quality of NBA data, select variables for predictive analysis,
apply machine learning techniques to predict NBA game results and evaluate the effectiveness of the
proposed methodology. The methodology includes data collection obtained through the data
scraping technique used on the official NBA website, data pre-processing, selection of the most
relevant variables through the PCA technique, modeling and evaluation. The results obtained
demonstrate an effective methodology for predictive analysis of results of sports games, regarding
the identification of the most important variables to predict results of NBA games and statistics of
prediction success. The present work, therefore, contributes to the advancement of predictive
analytics in sports and other fields of application.