Nonparametric (residual) kriging

Source: R/kriging.R

Compute simple kriging or residual kriging predictions (and also the corresponding simple kriging standard errors ). Currently, only global (residual) simple kriging is implemented.

np.kriging(object, ...)

# S3 method for default
np.kriging(object, svm, lp.resid = NULL, ngrid = object$grid$n, ...)

# S3 method for np.geo
np.kriging(object, ngrid = object$grid$n, ...)

kriging.simple(x, y, newx, svm)

kriging.simple.solve(x, newx, svm)

Arguments

object

object used to select a method: local polynomial estimate of the trend (class locpol.bin) or nonparametric geostatistical model (class extending np.geo).
...

further arguments passed to or from other methods.
svm

semivariogram model (of class extending svarmod).
lp.resid

residuals (defaults to residuals(object)).
ngrid

number of grid nodes in each dimension.
x

a (data) object used to select a method.
y

vector of data (response variable).
newx

vector/matrix with the (irregular) locations to predict (columns correspond with dimensions and rows with locations) or an object extending grid.par-class (data.grid).

See also

np.fitgeo, locpol, np.svar.