prep_par_output

(<code>numeric</code>) Found optimal value of
<code><a rd-options="stats" href="/link/optim?package=varycoef&version=0.3.3&to=stats" data-mini-rdoc="stats::optim">optim</a></code>.

output_par

(<code>spam</code> or <code>matrix(n, n)</code>) Covariance matrix
Sigma of SVC under final covariance parameters.

Sigma_final

(<code>NULL</code> or <code>spam.chol.NgPeyton</code>) If
covariance tapering is used, the Cholesky factor has been calculated
previously and can be used to efficiently update the Cholesky factor of
<code>Sigma_final</code>, which is an <code>spam</code> object.

Rstruct

(<code>logical(1)</code>) Indicates if optimization has been
conducted over full or profile likelihood.

profileLik

(<code>matrix(n, p)</code>) Design matrix

(<code>numeric(p)</code>) Response vector

(<code>NULL</code> or <code>matrix</code>) Hessian of MLE

(<code>numeric(1)</code>) Number of SVC

Prepares and computes the ML estimates and their respective standard errors.

Implements a maximum likelihood estimation (MLE)
method for estimation and prediction of Gaussian process-based
spatially varying coefficient (SVC) models
(Dambon et al. (2021a) <doi:10.1016/j.spasta.2020.100470>).
Covariance tapering (Furrer et al. (2006) <doi:10.1198/106186006X132178>) can
be applied such that the method scales to large data. Further, it implements
a joint variable selection of the fixed and random effects (Dambon et al.
(2021b) <arXiv:2101.01932>). The package and its capabilities are described
in (Dambon et al. (2021c) <arXiv:2106.02364>).

Jakob Dambon

varycoef

Modeling Spatially Varying Coefficients

Jakob A. Dambon

Fabio Sigrist

Reinhard Furrer

prep_par_output function

<dl><dt>output_par</dt>
<dd>(<code>numeric</code>) Found optimal value of
<code><a href="/link/optim?package=varycoef&version=0.3.3" data-mini-rdoc="varycoef::optim">optim</a></code>.</dd>
<dt>Sigma_final</dt>
<dd>(<code>spam</code> or <code>matrix(n, n)</code>) Covariance matrix
Sigma of SVC under final covariance parameters.</dd>
<dt>Rstruct</dt>
<dd>(<code>NULL</code> or <code>spam.chol.NgPeyton</code>) If
covariance tapering is used, the Cholesky factor has been calculated
previously and can be used to efficiently update the Cholesky factor of
<code>Sigma_final</code>, which is an <code>spam</code> object.</dd>
<dt>profileLik</dt>
<dd>(<code>logical(1)</code>) Indicates if optimization has been
conducted over full or profile likelihood.</dd>
<dt>X</dt>
<dd>(<code>matrix(n, p)</code>) Design matrix</dd>
<dt>y</dt>
<dd>(<code>numeric(p)</code>) Response vector</dd>
<dt>H</dt>
<dd>(<code>NULL</code> or <code>matrix</code>) Hessian of MLE</dd>
<dt>q</dt>
<dd>(<code>numeric(1)</code>) Number of SVC</dd></dl>

Arguments

Preparation of Parameter Output — prep_par_output

<dl>

<dt>output_par</dt>
<dd>(<code>numeric</code>) Found optimal value of
<code><a href='https://rdrr.io/r/stats/optim.html'>optim</a></code>.</dd>


<dt>Sigma_final</dt>
<dd>(<code>spam</code> or <code>matrix(n, n)</code>) Covariance matrix
Sigma of SVC under final covariance parameters.</dd>


<dt>Rstruct</dt>
<dd>(<code>NULL</code> or <code>spam.chol.NgPeyton</code>) If
covariance tapering is used, the Cholesky factor has been calculated
previously and can be used to efficiently update the Cholesky factor of
<code>Sigma_final</code>, which is an <code>spam</code> object.</dd>


<dt>profileLik</dt>
<dd>(<code>logical(1)</code>) Indicates if optimization has been
conducted over full or profile likelihood.</dd>


<dt>X</dt>
<dd>(<code>matrix(n, p)</code>) Design matrix</dd>


<dt>y</dt>
<dd>(<code>numeric(p)</code>) Response vector</dd>


<dt>H</dt>
<dd>(<code>NULL</code> or <code>matrix</code>) Hessian of MLE</dd>


<dt>q</dt>
<dd>(<code>numeric(1)</code>) Number of SVC</dd>

</dl>

prep_par_output: Preparation of Parameter Output

Description

Usage

Value

Arguments