Bayesian Melding Estimation of a Stochastic SEIR Model

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)One of the main problems in estimating stochastic SEIR models is that the data are not completely observed. In this case, the estimation is usually done by least squares or by MCMC. The Bayesian melding method is proposed to estimate SEIR models and to evaluate the likelihood in the presence of incomplete data. The method is illustrated by estimating a model for HIV/TB interaction in the population of a prison.172101111Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)Laboratorio EpifismaConselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP