Bayesian spatio‐temporal models based on discrete convolutions |
| |
| Authors: | Bruno Sans Ó Alexandra M Schmidt Aline A Nobre |
| |
| Institution: | 1. Department of Applied Mathematics and Statistics University of California 1156 High Street, Mail Stop: SOE2 Santa Cruz, CA 95064, USA;2. Department of Statistical Methods, IM‐UFRJ PO Box 68530, Rio de Janeiro, RJ Brazil 21.945–970;3. Scientific Computing Program, Oswaldo Cruz Foundation Av. Brasil, 4365 Manguinhos, Rio de Janeiro‐RJ Brazil 21045–900 |
| |
| Abstract: | Abstract: The authors consider a class of models for spatio‐temporal processes based on convolving independent processes with a discrete kernel that is represented by a lower triangular matrix. They study two families of models. In the first one, spatial Gaussian processes with isotropic correlations are convoluted with a kernel that provides temporal dependencies. In the second family, AR(p) processes are convoluted with a kernel providing spatial interactions. The covariance structures associated with these two families are quite rich. Their covariance functions that are stationary and separable in space and time as well as time dependent nonseparable and nonisotropic ones. |
| |
| Keywords: | Bayesian inference Gaussian process model comparison nonseparability non‐stationarity |
|
|
