Comparison of classic and Bayesian model for data on perinatal deaths at ISEA, Campina Grande-PB

• Autor(a) principal: Albuquerque, Mácio Augusto de
• Data de Publicação: 2020
• Outros Autores: Lucena, Sandro Lins Lopes de, Barros, Kleber Napoleão Nunes de Oliveira
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Research, Society and Development
• Texto Completo: https://rsdjournal.org/index.php/rsd/article/view/5477
• topic: Logistic regression; Bayesian inference; Perinatal death.; Regresión logística; Inferencia bayesiana; Muerte perinatal.; Regressão logística; Inferência Bayesiana; Óbito perinatal.

Generalized linear models are useful, among other situations, when you want to fit models to data that do not follow normality and cannot be adjusted using only simple linear regression. Another powerful estimation tool is the Bayesian methods, based on conditional probabilities. This work presents an adjustment of logistic regression models with parameters estimated by the maximum likelihood method, which is updated using Bayesian inference techniques. Such methods were applied to data obtained at the Instituto de Saúde Elpídio de Almeida, which is located in the city of Campina Grande - PB. The information refers to pregnant patients seen at this health unit. The objective was to obtain the best possible model that provides us with information about the chance of death of a child due to maternal variables using the maximum likelihood estimation method and the Bayesian method. The adjustments and diagnostics of the models were performed with the aid of the R software. It was found that the model estimated by the maximum likelihood is very close to the Bayesian model.