generalizedwendland-internal: Internal Functions defined in package 'GeneralizedWendland'
In GeneralizedWendland: Fully Parameterized Generalized Wendland Covariance Function

generalizedwendland-internal

R Documentation

Internal Functions defined in package GeneralizedWendland

Description

The functions or variables listed here are used internally, i.e. they are not intended to be called directly by the user.

Usage

control.wendland(numint.abstol = getOption("wendland.numint.abstol"),
  numint.reltol = getOption("wendland.numint.reltol"),
  numint.qag_key = getOption("wendland.numint.qag_key"),
  numint.subintervals = getOption("wendland.numint.subintervals"),
  interp.method = getOption("wendland.interp.method"),
  interp.num_support = getOption("wendland.interp.num_support"),
  cov.reparameterize = getOption("wendland.cov.reparameterize"),
  cov.eps = getOption("wendland.cov.eps"),
  cov.d_value = getOption("wendland.cov.d_value"), ...)

control.askey(cov.reparameterize = getOption("wendland.cov.reparameterize"),
  cov.eps = getOption("wendland.cov.eps"),
  cov.d_value = getOption("wendland.cov.d_value"), ...)

Arguments

`numint.abstol`	Absolute tolerance for numerical integration (default: 1e-3).
`numint.reltol`	Relative tolerance for numerical integration (default: 1e-3).
`numint.qag_key`	Method to use in QAG integration, from 0-6 (default: 0).
`numint.subintervals`	Number of subintervals to use in QAG/QAGS integration (default: 0).
`interp.method`	Method to use for covariance interpolation. Valid methods are 'none', 'linear', 'polynomial', and 'cspline' (default: 'none').
`interp.num_support`	Number of support points to use for covariance interpolation (default: 0).
`cov.reparameterize`	Whether to apply the reparameterization `\mu=\frac{1+d}{2} + \kappa + \nu`, where `\nu` takes the place of `\mu` in input vector `\vec{\theta}` (default: TRUE). This allows users to use box constraints in maximum likelihood estimation, as the covariance function is valid for `\nu \in [0,\infty)` rather than `\mu \in [\frac{1+d}{2} + \kappa,\infty)`.
`cov.eps`	The threshold distance `\epsilon` below which the function will return `\sigma + \theta` (default: .Machine$double.eps^0.5).
`cov.d_value`	Dimensionality of space in which measurements were taken (default: 2). This only takes effect if `cov.reparameterize` is TRUE.
`...`	Other arguments. Unused.

Details

The functions control.wendland and control.askey are helper functions which create a named list containing the relevant default arguments passed via the cov.args argument of cov.wendland and cov.askey respectively. In addition, users can override the default values.