Temos o prazer de convidar a todos para mais um seminário do Programa de Pós-Graduação em Estatística da UFRGS (PPGEst-UFRGS).
Título: Integrating Missing Data Handling into Probabilistic Regression Trees: The R Package PRTree.
Resumo: Probabilistic Regression Trees (PRTrees) extend traditional decision trees by incorporating probability functions that assign data points to multiple regions of the tree, enabling smoother decision boundaries and continuous predictions. In this work, we propose an adaptation of PRTrees that explicitly accounts for missing values in the covariates, enhancing its applicability to real-world, incomplete datasets. We implement this methodology in the R package PRTree, combining R with low-level FORTRAN and C routines to achieve substantial gains in computational efficiency. As a result, the package offers performance that is significantly superior to pure R implementations. In this presentation, we outline the theoretical foundations of the method, describe the implementation details, and illustrate its performance through simulation studies.