Bayesian Inference from Composite Likelihoods, with an Application to Spatial Extremes - Statistics > MethodologyReport as inadecuate




Bayesian Inference from Composite Likelihoods, with an Application to Spatial Extremes - Statistics > Methodology - Download this document for free, or read online. Document in PDF available to download.

Abstract: Composite likelihoods are increasingly used in applications where the fulllikelihood is analytically unknown or computationally prohibitive. Although themaximum composite likelihood estimator has frequentist properties akin to thoseof the usual maximum likelihood estimator, Bayesian inference based oncomposite likelihoods has yet to be explored. In this paper we investigate theuse of the Metropolis-Hastings algorithm to compute a pseudo-posteriordistribution based on the composite likelihood. Two methodologies for adjustingthe algorithm are presented and their performance on approximating the trueposterior distribution is investigated using simulated data sets and real dataon spatial extremes of rainfall.



Author: Mathieu Ribatet, Daniel Cooley, Anthony C. Davison

Source: https://arxiv.org/



DOWNLOAD PDF




Related documents