Correlated time series using mixed models in a Bayesian perspective
Roberto Molina de Souza Souza1; Jorge Alberto Achcar Achcar2; Glaucia Maria Bressan Bressan1
1 UTFPR; 2 USP
Resumo
The goal of this work is to propose the application of an autoregressive model considering two data series: hospital internment and inhalable particulate material with aerodynamic diameter less than 10 mm in a city in the state of São Paulo (oct/2003 to dez/2007). In a Bayesian perspective, using MCMC methods, we consider a mixed model with random effects to capture the possible correlation between the series and residuals with a Student-t distribution. This model is more appropriate when compared to independent models with normal residuals.
Palavras-chave: Time series correlated; Bayesian Inference; Student-t residuals; hospital internment; MCMC methods