reml: Restricted maximum likelihood (REML)

In qle: Simulation-Based Quasi-Likelihood Estimation

Description

Compute the value of the REML function (without constant term) for a given covariance model and data

Usage

reml(models, pars, data, Xs, verbose = FALSE)

Arguments

models	object of class krige (list of covariance models) or class covModel (a single covariance model), see setCovModel
pars	covariance parameter vector (including global scalar nugget value)
data	data frame of simulated statistics, each column corresponds to a single covariance model in 'models'
Xs	matrix of sample points
verbose	logical, if TRUE, print intermediate output

Details

Given a list of covariance models the function calculates the REML function values at the covariance parameter 'pars'.

Value

List of REML function values.

Author(s)

M. Baaske

Examples

	data(normal)

 # extract the sample points (the design)
 X <- as.matrix(qsd$qldata[1:2])

 # get the observed statistic
 T <- qsd$qldata[c("mean.T1")]
 reml(qsd$covT[1],pars=c(1e-4,1.5,0.1),T,X)

qle documentation built on May 2, 2019, 9:55 a.m.