mleFactory {GeneralizedWendland}

Function Factory for Generating mle Function with unified arguments

Description

A factory function which returns a function of the form function(y, X = data.frame(), distmat, init_parameters, theta_llim, theta_ulim) which can be called to compute the maximum likelihood estimates for a Kriging model.

Usage

mleFactory(covariance, cov.args = list(), chol.args = list(),
  optim.args = list(), hessian = FALSE, optimParallel.args = list())

Arguments

covariance

A function of the form function(h, theta, ..., cov.args = list()), where h is an object storing a distance matrix, theta is a numeric vector of parameters for the linear predictor and covariance function, and cov.args is a list of optional arguments for the covariance function.

cov.args

A list of optional settings for a covariance function.

chol.args

A list of optional settings for a cholesky function. Note: Valid input arguments change depending on whether the distance matrix provided to the output function is sparse. This may change in a future version.

optim.args

A list of optional settings for optim. See optim for documentation of valid arguments.

hessian

A logical value which specifies whether the hessian matrix is to be returned in the output. Is FALSE by default.

optimParallel.args

A list of optional settings for optimParallel. See optimParallel for documentation of valid arguments.

Details

The purpose of this function factory is to return an mle function with unified arguments. The returned function performs the same task as for example spam::mle(), but simplifies the process in two ways: The returned function detects whether the Gaussian process is a zero-mean process through the input argument X and whether methods from the spam package should be used based on the type of input argument distmat, and autonomously chooses appropriate methods to compute the neg2loglikelihood. Hence the user does not need to choose a specialized method themselves.

Value

A function of the form function(y, X = data.frame(), distmat, beta0 = NULL, init_parameters, theta_llim, theta_ulim) which returns the output of optim or optimParallel if optimParallel.args was specified.

Author(s)

Thomas Caspar Fischer

References

Hadley Wickham (2015) Advanced R, CRC Press.

See Also

optim, optimParallel, covarianceFactory, choleskyFactory and optimFactory

Examples

set.seed(57)
n <- 50
range <- 0.4
theta  <- c(range, 1, 1, 0, 0)

locs <- data.frame(x = runif(n), y = runif(n))
dmat  <- as.matrix(dist(locs))
Sigma <- cov.wendland(h = dmat, theta = theta)
y <- c(spam::rmvnorm(1, Sigma = Sigma))

init_parameters   <- c(0.7, 2, 0, 2, 2)
lower_constraints <- c(0.1, 0.1, 0, 0, 0)
upper_constraints <- c(sqrt(2), 2, 2, 2, 2)

mleFunction <- mleFactory(covariance = cov.wendland)
mle_result1 <- mleFunction(y = y, distmat = dmat,
                           init_parameters = init_parameters, theta_llim = lower_constraints,
                           theta_ulim = upper_constraints)

mleFunctionDM <- mleFactory(covariance = cov.wendland,
                            cov.args = list(fixed_range_value = range))
mle_result2 <- mleFunctionDM(y = y, X = data.frame(), distmat = dmat,
                             init_parameters = init_parameters[-1],
                             theta_llim = lower_constraints[-1],
                             theta_ulim = upper_constraints[-1])