作者: Andreas Hofinger , Hanna K Pikkarainen
DOI: 10.1088/0266-5611/23/6/012
关键词: Prior probability 、 Inverse problem 、 Posterior probability 、 Identity matrix 、 Rate of convergence 、 Mathematical optimization 、 Applied mathematics 、 Mathematics 、 Convergence tests 、 Covariance 、 Bayesian linear regression
摘要: Recently, the metrics of Ky Fan and Prokhorov were introduced as a tool for studying convergence regularization methods stochastic ill-posed problems. In this work, we show that Bayesian approach to linear inverse problems can be examined in new framework well. We consider finite-dimensional case where measurements are disturbed by an additive normal noise, prior distribution is normal. Convergence rate results posterior obtained when covariance matrices proportional identity matrix.