Quotation Frühwirth-Schnatter, Sylvia. 2003. Parsimonious Markov Chain Monte Carlo for Models with Latent Variables. Joint Statistical Meetings 2003, San Francisco, Vereinigte Staaten/USA, 03.08.-07.08.




A well-known problem for complex models involving latent variables is slow convergence of straightforward MCMC schemes. A common cause is poor parameterization of the latent structure in combination with models that are overparameterized in light of the data. A typical example would be a random-effects model with some of the random effects being nearly deterministic. We consider parsimonious MCMC methods where sampling the unknown model parameters is carried out jointly with finding a parsimonious representation of the underlying model structure. To this aim, suitable selection variables are introduced that are sampled jointly with the parameters. Details will be presented for hierarchal random-effects models. Model flexibility is introduced both for the mean structure, as in common Bayesian variable selection, as well as in the covariance structure of the latent process. Joint covariance selection in combination with choosing the right parameterization of the latent process is shown to be an efficient computational tool for carrying out MCMC.


Press 'enter' for creating the tag

Publication's profile

Status of publication Published
Affiliation WU
Type of publication Paper presented at an academic conference or symposium
Language English
Title Parsimonious Markov Chain Monte Carlo for Models with Latent Variables
Event Joint Statistical Meetings 2003
Year 2003
Date 03.08.-07.08
Country United States/USA
Location San Francisco
URL http://www.amstat.org/meetings/jsm/2003/onlineprogram/index.cfm?fuseaction=abstract_details&abstractid=301718


Frühwirth-Schnatter, Sylvia (Details)
Institute for Statistics and Mathematics IN (Details)
Google Scholar: Search