Gibbs sampling on a steady model

Abstract

In this study Gibbs sampling, a widely used simulation method, is applied to the steady model, a simple variation of the dynamic linear model, and the model parameters are estimated. The estimates obtained from Gibbs sampling and the results for the standard Kalman filter are compared and are found to be close. These similarities in the results indicate the success of the stochastic simulation. In this study, a variance modulation on the steady model is also applied and Gibbs sampling is proposed to overcome analytic problems. In the variance adaptation, denned as $mu^b_t(a,b>0)$, estimates for the model parameters are obtained for different values of a and b.