The Study on MCMC Methods of Nonlinear Dynamic Models

Abstract: In this paper, firstly we introduce several kinds of random simulation methods that are often used. Then we infer and forecast the Nonlinear Bayesian Dynamic Models with rejection acceptance algorithm and resample move algorithm. Because there are a lot of unknown parameters in Nonlinear Dynamic Models, it still is difficult to estimate parameters. Monte Carlo optimization is used to estimate parameters. This method is also used in ARCH(0,p) models to estimate parameters. Finally, we discuss the approaches of Bayesian model selection. A new method is introduced to select models. We place the problem within a decision theoretic framework…
Key words: Nonlinear Dynamic Models; Random simulation; Parameter estimation; Model selection

This entry was posted in Master Thesis. Bookmark the permalink.