Classifcação binária via Bayes Ingênuo: um estudo comparativo de predições

Abstract

In this undergraduate thesis, we propose a review of the Naive Bayes classifcation method applied to binary response variables, with a more in-depth formalization of the Gaussian Naive Bayes and Flexible Naive Bayes methods for numerical covariates. To compare these methods using the statistical software R, we will use a database containing both numerical and categorical covariates, with the binary response variable representing victory or defeat.