Please use this identifier to cite or link to this item: http://repositorio.ufla.br/jspui/handle/1/12433
Title: Modelo beta inflacionado de zeros e uns: análise bayesiana de notas no curso de estatística aplicada à engenharia
Authors: Bueno Filho, Júlio Silvio de Souza
Cirillo, Marcelo Angelo
Nogueira, Denismar Alves
Keywords: Distribuição Beta Inflacionada de Zeros e Uns
Inferência Bayesiana
Métodos Monte Carlo via Cadeias de Markov (MCCM)
Bayesian Inference
Zero and One Inflated Beta Distribution
Estatística
Issue Date: 7-Mar-2017
Publisher: Universidade Federal de Lavras
Citation: FREITAS JÚNIOR, H. da S. Modelo beta inflacionado de zeros e uns: análise bayesiana de notas no curso de estatística aplicada à engenharia. 2017. 92 p. Dissertação (Mestrado em Estatística e Experimentação Agropecuária)-Universidade Federal de Lavras, Lavras, 2017.
Abstract: This dissertation discusses an implementation of Bayesian inference on parameters of a Zero and One Inflated Beta Distribution (zoib). Data used to illustrate the proceeding were grades in Applied Statistics from Engeneering Students in second semester of 2015. Algorithms were tested in a simmulation study with different sample sizes and combinaions of parametric values. This study has shown routines are unbiased and consistent. Samples from the joint posterior distribution were drawn using Gibbs Sampler with a taylor made Gaussian random walk to yield better sampling properties for the full conditional distributions of parameters of the continuous part of the model (m and f ). The actual data analysis was carried out in 13 out of the 14 items with zeroes and ones inflation. Analysis can be done routinely to examples alike and can generate usefull insight for researchers in the applied area (educational testing).
URI: http://repositorio.ufla.br/jspui/handle/1/12433
Appears in Collections:Estatística e Experimentação Agropecuária - Mestrado (Dissertações)



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.