rr_lognormal_krige: Spatial predictions for lognormal process

In jelsema/RRSM: Methods for reduced rank spatial models

Description

Obtains spatial predictions at selected locations for a lognormal process.

Usage

rr_lognormal_krige(Y, X, S = NULL, coords, pgrid = coords, Xpred = X,
  Spred = S, V = NULL, ssq, tsq = 0, ncore = 1, ...)

Arguments

Y	the observed data. Should already be on log-scale.
X	the design matrix for the large-scale variation (fixed effects).
S	the matrix of basis functions (design matrix for the random effect).
coords	the matrix of observed locations.
pgrid	the prediction grid.
Xpred	the design matrix for the predicted locations.
Spred	the matrix of basis functions for the predicted locations.
V	the estimated covariance matrix.
ssq	the estimated residual variance.
tsq	the estimated variance due to dimension reduction.
ncore	if ncore>1, number of cores to use with doParallel backend.
...	space for additional arguments.

Value

A list containing [[1]] a matrix with the coordinates of the predicted locations, the predicted value, and measures of uncertainty, and [[2]] the estimate of large-scale variation.

Note

The parameter estimates may be left empty, and ... may be used to pass arguments to estimation functions.

jelsema/RRSM documentation built on May 19, 2019, 4:02 a.m.