Applied Bayesian Data Analysis Using State-Space Models

Abstract

This paper reviews the Bayesian approach to parameter estimation in nonlinear nonnormal state-space models with posterior computations performed by Gibbs sampling. Fitting of nonlinear non normal state-space models is an important task in various scientific disciplines. The ease with which the Bayesian approach can now be implemented via BUGS, a recently developed, user-friendly, and freely available software package, should have a major impact on applied research. This is illustrated using examples from three different areas of currently active research: econonometrics, fisheries, and physics.