Bayesian analysis of regression models with spatially correlated errors and missing observations

Man Suk Oh, Dong Wan Shin, Han Joon Kim

Research output: Contribution to journalArticlepeer-review

9 Scopus citations

Abstract

A Bayesian approach is proposed for estimating regression models on rectangular grids in which errors are spatially correlated and missing observations are present in the response variable. An easy and efficient Markov chain Monte Carlo algorithm is fully described for posterior inference on parameters and prediction of missing observations. Analysis of a real marine remote-sensing data set is presented to illustrate the method.

Original languageEnglish
Pages (from-to)387-400
Number of pages14
JournalComputational Statistics and Data Analysis
Volume39
Issue number4
DOIs
StatePublished - 28 Jun 2002

Bibliographical note

Funding Information:
The authors are very grateful for the valuable comments of the three referees and the previous Editor, Dr. Naeve. Also, we are indebted to Professor Paul Green for the careful proofreading of this paper. The first author was supported by 1997 Ewha Womans University Research Fund.

Keywords

  • Gibbs sampling algorithm
  • Markov chain Monte Carlo
  • Missing value
  • Posterior inference
  • Spatial data

Fingerprint

Dive into the research topics of 'Bayesian analysis of regression models with spatially correlated errors and missing observations'. Together they form a unique fingerprint.

Cite this