Bayesian Analysis of Non-Stationary Gaussian Process Models
Calculate the Gaussian quadratic form for the NNGP approximation
Calculate A and D matrices for the NNGP approximation
Calculate the (sparse) matrix U
Assign conditioning sets for the SGV approximation
Calculate sparse kernel, core kernel, and determine nonzero entries
Calculate sparse kernel, core kernel, and determine nonzero entries
Determine the k-nearest neighbors for each spatial coordinate.
Function for the evaluating the Gaussian likelihood with gp2Scale spar...
Function for the evaluating the NNGP approximate density.
Function for the evaluating the SGV approximate density.
Calculate covariance elements based on eigendecomposition components
Calculate a stationary Matern correlation matrix
nimble_sparse_chol
nimble_sparse_chol
nimble_sparse_crossprod
nimble_sparse_solve
nimble_sparse_crossprod
nimble_sparse_tcrossprod
Calculate a nonstationary Matern correlation matrix
Calculate a nonstationary Matern cross-correlation matrix
Calculate coordinate-specific cross-distance matrices
Calculate coordinate-specific cross-distance matrices, only for neares...
Calculate coordinate-specific distance matrices
Calculate coordinate-specific distance matrices, only for nearest neig...
NIMBLE code for a generic nonstationary GP model
Posterior prediction for the NSGP
Order coordinates according to a maximum-minimum distance criterion.
R_sparse_chol
R_sparse_chol
nimble_sparse_crossprod
nimble_sparse_solve
nimble_sparse_crossprod
nimble_sparse_tcrossprod
Function for the evaluating the SGV approximate density.
Function for the evaluating the NNGP approximate density.
Function for the evaluating the SGV approximate density.
One-time setup wrapper function for the SGV approximation
Enables off-the-shelf functionality for fully Bayesian, nonstationary Gaussian process modeling. The approach to nonstationary modeling involves a closed-form, convolution-based covariance function with spatially-varying parameters; these parameter processes can be specified either deterministically (using covariates or basis functions) or stochastically (using approximate Gaussian processes). Stationary Gaussian processes are a special case of our methodology, and we furthermore implement approximate Gaussian process inference to account for very large spatial data sets (Finley, et al (2017) <doi:10.48550/arXiv.1702.00434>). Bayesian inference is carried out using Markov chain Monte Carlo methods via the "nimble" package, and posterior prediction for the Gaussian process at unobserved locations is provided as a post-processing step.