Bayesian Inference of Spatial Covariance Parameters |
| |
| Authors: | Eulogio Pardo-Igúzquiza |
| |
| Institution: | (1) TAMSAT, Department of Meteorology, The University of Reading, P.O. Box 243, Earley Gate, Reading, RG6 6BB, United Kingdom |
| |
| Abstract: | This paper shows the application of the Bayesian inference approach in estimating spatial covariance parameters. This methodology is particularly valuable where the number of experimental data is small, as occurs frequently in modeling reservoirs in petroleum engineering or when dealing with hydrodynamic variables in groundwater hydrology. There are two main advantages of Bayesian estimation: firstly that the complete distribution of the parameters is estimated and, from this distribution, it is a straightforward procedure to obtain point estimates, confidence regions, and interval estimates; secondly, all the prior information about the parameters (information available before the data are collected) is included in the inference procedure through their prior distribution. The results obtained from simulation studies are discussed. |
| |
| Keywords: | geostatistics prior distribution likelihood function posterior distribution variogram covariance range |
| 本文献已被 SpringerLink 等数据库收录！ |
|
