check.model {RGeostats} | R Documentation |
Checking that the Model is valid
check.model(model, nvar.func = 1, ndim.func = 2, must.not.be.func = FALSE)
model |
This contains the argument that should be checked against a 'Model'. Two modes are available:
|
nvar.func |
When the Model is provided as an external function, this argument provides the number of variables for which the function is defined |
ndim.func |
When the Model is provided as an external function, this argument provides the space dimension for which the function is defined |
must.not.be.func |
When TRUE, the 'Model' argument may not be given as a function. |
When the 'Model' is provided as a function, the prototype of this function is given as:
my_cov_func(dist,incr=NA,x1=NA,x2=NA,...)
where
dist: Euclidean distance between the two end-points
db1: Rank of the Db at first point (1 for Dbin; 2 for dbout)
iech1: Rank of sample within first Db at first point (starts at 1)
db2: Rank of the Db at second point (1 for Dbin; 2 for dbout)
iech2: Rank of sample within second Db at second point (starts at 1)
incr: Vector giving the distance increments between the two end-points (Dimension: Space dimension)
x1: Vector describing the coordinates of the first end-point. This argument is provided only in the non-stationary case
x2: Vector describing the coordinates of the second end-point. This argument is provided only in the non-stationary case
In the external function, arguments may be missing (except 'dist'). For security, the argument '...' is recommanded. Moreover, the arguments 'db1', 'iech1', 'db2' and 'iech2' are only defined when the external covariance function is called within a Kriging procedure. Otherise, these arguments are returned as 0 (for 'db*") and -1 (for iech*).
This function returns the following information:
nvar: Number of variables
ndim: Space dimension
ncova: Number of basic structures