Study on Forecasting and Simulation of the Nonlinear Bayesian Dynamic Models

Abstract: In this paper, I discuss mainly the using of the simulation in the non-linear Bayesian Dynamic Models. Regard to the two kinds non-linear Bayesian Dynamic Models, I give the simulation of their own. For example, to the general models such as follows:With the application of the Serial Important Function and the Gibbs Algorithms in MCMC methods, I give some results in the choice of the important function and how to use the Gibbs Algorithms in the forecasting of the models. To the models as follows:With θ in zt(θ).With the application of the M. West's Kemal theory and the Sequential imputation Algorithm, we can solve the problem of the forecasting. At the same time, I discover the relations between the M. West method and Gonden et al method in the discussing of the information losing…
Key words: the non-linear Bayesian Dynamic Models; the Serial Important Sample; the Serial Imputation Algorithm; the MCMC methods; M.West method; Gonden et al method; The Bayesian Factor

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