Geweke, J. and H. Tanizaki (1999). "On Markov Chain Monte Carlo Methods for Nonlin- ear and Non-Gaussian State-Space Models". Communications in Statistics, Simulation and Computation, 28, pp. 867-894.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....of samples at each step of the update. It does, however, suffer from the phenomenon of sample impoverishment , which in the statistical literature is treated with a varietyof remedies, including the Rao Blackwellization procedure [10] 2. 3 Other sampling strategies GewekeandTanizaki [7] describe a Monte Carlo algorithm based on Rejection Sampling for the filtering density ae(x t ) p(x t jY t ) ae(x t 1 ) kp(y t 1 jx t 1 ) Z p(x t 1 jx t )p(x t jY t )d(x t ) 27) which is based upon drawing samples from p(x t 1 jx t ) and accepting them with probability proportional ....

J. Geweke and H. Tanizaki. On markovchain monte carlo methods for nonlinear and non gaussian state-space models. commun. Stat. Simul. C, 28:867--894, 1999.

....acceptance probability is close to zero, rejection sampling takes a long time computationally. Moreover, we have the case where the acceptance probability is zero. In such a case, rejection sampling cannot be applied. In order to avoid these computational disadvantages of the existing procedures, Geweke and Tanizaki (1999) suggested the nonlinear and or non Gaussian smoother applying the Metropolis Hastings algorithm and the Gibbs sampler simultaneously, where the measurement and transition equations are specified in any general formulation and the error terms in the state space model are not necessarily normal. ....

....framework. Utilizing the Metropolis Hastings algorithm in addition to the Gibbs sampler, in this paper, we deal with any nonlinear and or non Gaussian state space model in a Bayesian framework. Thus, this paper is an extension of Carlin, Polson and Sto#er (1992) Carter and Kohn (1994, 1996) and Geweke and Tanizaki (1999). 2 Moreover, several proposal densities are taken and compared by some Monte Carlo studies, since the critical problem of the Metropolis Hastings algorithm is choice of the proposal density. We conclude by the root mean square criterion that the proposed procedure are not a#ected by choice of ....

[Article contains additional citation context not shown here]

Geweke, J. and Tanizaki, H., 1999, " On Markov Chain Monte Carlo Methods for Nonlinear and Non-Gaussian State-Space Models, " Communications in Statistics, Simulation and Computation, Vol.28, No.4, pp.867 -- 894.

No context found.

Geweke, J. and Tanizaki, H., 1999a, " On Markov Chain Monte Carlo Methods for Nonlinear and NonGaussian State-Space Models, " Communications in Statistics, Simulation and Computation, Vol.28, No.4, pp.867 -- 894.

....sampling, rejection sampling (RS) importance resampling (IR) the MetropolisHastings independence sampling (MH) and etc. Carlin et al. 1992) and Carter and Kohn (1994, 1996) applied the Gibbs sampler to some specific state space models, which are extended to more general state space models by Geweke and Tanizaki (1999). The Gibbs sampler sometimes gives us the imprecise estimates of the state variables, depending on the underlying state space model (see Carter and Kohn (1994, 1996) Especially when the random variables are highly correlated with each other, it is well known that convergence of the Gibbs ....

Geweke, J. and Tanizaki, H. (1999). On Markov Chain Monte-Carlo Methods for Nonlinear and NonGaussian State-Space Models, Communications in Statistics, Simulation and Computation, 28, 867--894.

No context found.

Geweke, J. and H. Tanizaki (1999). "On Markov Chain Monte Carlo Methods for Nonlin- ear and Non-Gaussian State-Space Models". Communications in Statistics, Simulation and Computation, 28, pp. 867-894.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC