RandomFields (version 3.1.16)

RFgetModelInfo: Information on RMmodels

Description

The function returns information about a RMmodel, either internal information when used in simulations, for instance, or general information

Usage

RFgetModelInfo(...) 
RFgetModelInfo_register(register, level = 1, spConform = RFoptions()$general$spConform, which.submodels = c("user", "internal", "call+user", "call+internal", "user.but.once", "internal.but.once", "user.but.once+jump", "internal.but.once+jump", "all"), modelname = NULL)
RFgetModelInfo_model(model, dim = 1, Time = FALSE, kernel = FALSE, exclude_trend = TRUE, ...)

Arguments

...

any of the arguments below

register

$0,...,21$ or an evaluating function, e.g. RFsimulate. Place where intermediate calculations are stored. See also section Registers in RFoptions.

level

integer [0..5]; level of details, i.e. the higher the number the more details are given.

spConform

see RFoptions

which.submodels

Internally, the sub-models are represented in two different ways ‘internal’ and ‘user’. The latter is very close to the model defined by the user. Most models have a leading internal model. The values "call+user", "call+internal" return also this leading model if existent. The values "user.but.once", "internal.but.once" "user.but.once" returns the user path of the internal model following the leading model. "internal.but.once" would return the internal path of the user model following the leading model, but this path should never exist. So as all the other options if a certain direction does not exist, the alternative path is taken. The values "user.but.once+jump", "internal.but.once+jump" same as "user.but.once" and "internal.but.once", except that the first submodel below the leading model is not given.

The value "all" returns the whole tree of models (very advanced).

modelname

string. If modelname is given then it returns the first appearance of the covariance model with name modelname. If meth is given then the model within the method is returned.

model

an RMmodel with NAs where information on the parameters is requested.

dim

positive integer. Spatial dimension.

Time

logical. Should time be considered too?

kernel

logical. Should the model be considered as a kernel?

exclude_trend

logical. Currently, only TRUE is available.

Value

If RFgetModelInfo(model) is called a list returned with the following elements:

• trans.inv : logical. Whether the model is translation invariant (stationary)
• isotropic : logical. Whether the model is rotation invariant (stationary)
• NAs : is case of an additive model it gives the number of NAs in each submodel
• minmax : a data frame containing information on all arguments set to NAs
  • pmin, pmax : lower and upper endpoint of the parameter values usually found in practice
  • type : integer; recognized particularities of a parameter; an explanation of the values is given after the table, if printed.
  • NAN : the number of NANs found
  • min, max : mathetically valid lower and upper endpoints of the parameter values
  • omin, omax : logical. If FALSE the respective mathematical endpoint is included
  • col, row : the dimension of the parameter. If the parameter is a scalar then col = row = 1. If it is a vector then col = 1.
  • bayes : currently not used (always FALSE)

Else a list of internal structure is returned.

Details

RFgetModelInfo branches either into RFgetModelInfo_register or RFgetModelInfo_model, depending on the type of the first argument. The latter two are usually not called by the user.

RFgetModelInfo has three standard usages:

• RFgetModelInfo() returns internal information on the last call of an RF function.
• RFgetModelInfo(RFfunction) returns internal information on the last call of RFfunction.
• RFgetModelInfo(RMmodel) returns general information on RMmodel

Whereas RFgetModelInfo() can return detailed internal information, RFgetModel returns a model that can be re-used by the user.

See Also

commandRFgetModel, RFsimulate

Examples

RFoptions(seed=0) ## *ANY* simulation will have the random seed 0; set
##                   RFoptions(seed=NA) to make them all random again

model <- RMexp(scale=4, var=2) + RMnugget(var=3) + RMtrend(mean=1)
z <- RFsimulate(model, 1:4, storing=TRUE)
RFgetModelInfo()

model <-  RMwhittle(scale=NA, var=NA, nu=NA) + RMnugget(var=NA)
RFgetModelInfo(model)