Bayesian Spatial Prediction for Closed Skew Gaussian Random Field

摘要

When the distribution of a data set is skewed but it has many similar properties as normal distribution, skew-normal distribution can be used to model their skewness. A large class of multivariate skew-normal distribution was introduced, namely "closed skew-normal distribution", which has the advantage of being closed under marginalization and conditioning. In this work, we tried to generalize die prediction method for a closed skew Gaussian random field to provide a Bayesian spatial predictor. Then a simulation study is performed to check validity of the model and performance of the Bayesian spatial predictor.