Regressão linear múltipla na modelagem de resultados na National Basketball Association (NBA)

Abstract

Sports in general play a very important role in society. Its benefits are realized both in terms of physical and mental health, as well as entertainment, and very important economically for a region. Given this fact, the use of appropriate statistical techniques, both descriptive and decision-making becomes an important tool for planning actions that enable better results. In this context, this work evaluated, through the descriptive statistics of the collected information and the method of multiple linear regression analysis, which regressive variables, statistics of the basketball game, were significant to explain the response variable, total number of team wins. of the professional basketball league in the United States, the National Basketball Association (NBA). Final model Showing a good fit, being able to explain 94 % of the variability found in the total number of wins. To perform the analyzes, data were collected on the oficial NBA website at the end of 9 regular seasons (2009-2019). All analyzes and results were performed and obtained using software R.