A copula based approach to adaptive sampling - Statistics > MethodologyReport as inadecuate

A copula based approach to adaptive sampling - Statistics > Methodology - Download this document for free, or read online. Document in PDF available to download.

Abstract: Our article is concerned with adaptive sampling schemes for Bayesianinference that update the proposal densities using previous iterates. Weintroduce a copula based proposal density which is made more efficient bycombining it with antithetic variable sampling. We compare the copula basedproposal to an adaptive proposal density based on a multivariate mixture ofnormals and an adaptive random walk Metropolis proposal. We also introduce arefinement of the random walk proposal which performs better for multimodaltarget distributions. We compare the sampling schemes using challenging butrealistic models and priors applied to real data examples. The results showthat for the examples studied, the adaptive independent \MH{} proposals aremuch more efficient than the adaptive random walk proposals and that in generalthe copula based proposal has the best acceptance rates and lowestinefficiencies.

Author: Ralph Silva, Robert Kohn, Paolo Giordani, Xiuyan Mun

Source: https://arxiv.org/

Related documents