Bayesian spatial prediction for extended skew-t random field

Document Type : Original Article

Authors
Mehrdad Taghipour
Mohammad Mehdi Saber
Eghlid Center of Higher Education
10.22034/mmr.2026.11730
Abstract
This article introduces a random field defined by the generalized t-squared multivariate distribution، which allows for improved accuracy in spatial prediction. With its flexibility for heavy-tailed and light-tailed data، this random field is suitable for predicting both types of spatial data. The simulation results demonstrate that incorporating this random field enhances prediction accuracy.، particularly in scenarios with limited spatial data.
Keywords
spatial data
Gaussian random field
Multivariate extended skew-t distributions
spatial Bayesian prediction