corLevDiag

Compute the correlation or covariance matrix for a diagonal
 structure.

Gaussian process regression with an emphasis on kernels.
Quantitative and qualitative inputs are accepted. Some pre-defined
kernels are available, such as radial or tensor-sum for
quantitative inputs, and compound symmetry, low rank, group kernel
for qualitative inputs. The user can define new kernels and
composite kernels through a formula mechanism. Useful methods
include parameter estimation by maximum likelihood, simulation,
prediction and leave-one-out validation.

Olivier Roustant

kergp

Gaussian Process Laboratory

corLevDiag function

<dl> 
 <dt>par</dt>
<dd></dd>A numeric vector with length <code>npVar</code> where
 <code>npVar</code> is the number of variance parameters, namely
 <code>0</code>, <code>1</code> or <code>nlevels</code> corresponding to the values
 of <code>cov</code>: <code>0</code>, <code>1</code> and <code>2</code>.
 <dt>nlevels</dt>
<dd>
 
Number of levels.
</dd> <dt>levels</dt>
<dd></dd>Character representing the levels.
 <dt>lowerSQRT</dt>
<dd></dd>Logical. When <code>TRUE</code> the (lower) Cholesky
 root \(\mathbf{L}\) of the correlation or covariance matrix
 \(\mathbf{C}\) is returned instead of the correlation matrix.
 <dt>compGrad</dt>
<dd></dd>Logical. Should the gradient be computed?
 <dt>cov</dt>
<dd></dd>Integer <code>0</code>, <code>1</code> or <code>2</code>. If <code>cov</code>
 is <code>0</code>, the matrix is a correlation matrix (or its
 Cholesky root) i.e. an identity matrix. If <code>cov</code> is <code>1</code>
 or <code>2</code>, the matrix is a covariance (or its square
 root) with constant variance vector for <code>code = 1</code> and with
 arbitrary variance vector for <code>code = 2</code>.
</dl>

Arguments

Compute the correlation or covariance matrix for a diagonal
 structure.

Correlation or Covariance Matrix for a Diagonal Structure — corLevDiag

<dl>

 
 <dt>par</dt>
<dd>

A numeric vector with length <code>npVar</code> where
 <code>npVar</code> is the number of variance parameters, namely
 <code>0</code>, <code>1</code> or <code>nlevels</code> corresponding to the values
 of <code>cov</code>: <code>0</code>, <code>1</code> and <code>2</code>.
</dd>

 <dt>nlevels</dt>
<dd>
 
Number of levels.
</dd>

 <dt>levels</dt>
<dd>

Character representing the levels.
</dd>

 <dt>lowerSQRT</dt>
<dd>

Logical. When <code>TRUE</code> the (lower) Cholesky
 root \(\mathbf{L}\) of the correlation or covariance matrix
 \(\mathbf{C}\) is returned instead of the correlation matrix.
</dd>

 <dt>compGrad</dt>
<dd>

Logical. Should the gradient be computed?
</dd>

 <dt>cov</dt>
<dd>

Integer <code>0</code>, <code>1</code> or <code>2</code>. If <code>cov</code>
 is <code>0</code>, the matrix is a correlation matrix (or its
 Cholesky root) i.e. an identity matrix. If <code>cov</code> is <code>1</code>
 or <code>2</code>, the matrix is a covariance (or its square
 root) with constant variance vector for <code>code = 1</code> and with
 arbitrary variance vector for <code>code = 2</code>.
</dd>

</dl>

corLevDiag: Correlation or Covariance Matrix for a Diagonal Structure

Description

Usage

Value

Arguments

Examples