期刊
INVERSE PROBLEMS AND IMAGING
卷 13, 期 1, 页码 1-29出版社
AMER INST MATHEMATICAL SCIENCES-AIMS
DOI: 10.3934/ipi.2019001
关键词
Bayesian statistical estimation; inverse problems; Matern fields; hypermodels; convergence
资金
- Engineering and Physical Sciences Research Council, United Kingdom (EPSRC) [EP/K034154/1]
- Academy of Finland [250215, 307741, 313709]
- Academy of Finland (AKA) [307741, 313709, 307741, 313709] Funding Source: Academy of Finland (AKA)
We introduce non-stationary Matern field priors with stochastic partial differential equations, and construct correlation length-scaling with hyperpriors. We model both the hyperprior and the Matern prior as continuous-parameter random fields. As hypermodels, we use Cauchy and Gaussian random fields, which we map suitably to a desired correlation length-scaling range. For computations, we discretise the models with finite difference methods. We consider the convergence of the discretised prior and posterior to the discretisation limit. We apply the developed methodology to certain interpolation, numerical differentiation and deconvolution problems, and show numerically that we can make Bayesian inversion which promotes competing constraints of smoothness and edge-preservation. For computing the conditional mean estimator of the posterior distribution, we use a combination of Gibbs and Metropolis-within-Gibbs sampling algorithms.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据