ENFOQUE DEL MODELO DE REGRESIÓN LOGÍSTICA USANDO EL MUESTREADOR DE GIBBS

Abstract

The logistic regression model is very useful in different areas of the science for studying forecast problems when the response variable is binary. This model might be studied by the Bayesian approach, but in this case the computations can be very difficult. The Gibbs Sampler algorithm supported on the Winbugs software is an alternative method to obtain of posterior distribution of the parameters of the model. This software does not show how the results are obtained. For this reason in this work a new implementation in MATLAB of this algorithm is proposed. A comparison of both implementation was made through a bioclimatic forecast example of the frequency of the ASMA disease and a simulation study