G. Garcia-Donato, M. E. Castellanos Nueda
In this talk we introduce an exact algorithm that allows the exact implementation of the Gibbs algorithm for variable selection when the number of covariates is large or very large. We illustrate its performance in real datasets. We also introduce a new artificial experiment based on real data and sensible arguments. It has also the property that inclusion probabilities can be exactly calculated making it possible to evaluate the performance of numerical algorithms.
Palabras clave: Bayes factors, Large p, Multiplicity, Parallel computing
Programado
GT11.BAYES1 Selección de Variables
7 de noviembre de 2023 16:50
CC1: Auditorio