Matern32Kernel Class Reference

Stationary kernel with C^1 smooth realizations. More...

Inheritance diagram for Matern32Kernel:

Public Member Functions
void	compute_gram (const std::vector< MatrixXd > &dists2, const VectorXd &theta_values, MatrixXd &gram) override
	Compute a Gram matrix given a vector of squared distances and kernel hyperparameters. More...
void	compute_gram_derivs (const MatrixXd &gram, const std::vector< MatrixXd > &dists2, const VectorXd &theta_values, std::vector< MatrixXd > &gram_derivs) override
	Compute the derivatives of the Gram matrix with respect to the kernel hyperparameters. More...
MatrixXd	compute_first_deriv_pred_gram (const MatrixXd &pred_gram, const std::vector< MatrixXd > &mixed_dists, const VectorXd &theta_values, const int index) override
	Compute the first derivatve of the prediction matrix for a given component. More...
MatrixXd	compute_second_deriv_pred_gram (const MatrixXd &pred_gram, const std::vector< MatrixXd > &mixed_dists, const VectorXd &theta_values, const int index_i, const int index_j) override
	Compute the second derivatve of the prediction matrix for a pair of components. More...

Private Attributes
const double	sqrt3 = sqrt(3.)

Additional Inherited Members
Protected Member Functions inherited from Kernel
void	compute_Dbar (const std::vector< MatrixXd > &cw_dists2, const VectorXd &theta_values, bool take_sqrt=true)
	Compute the `‘Dbar’' matrices of scaled distances. More...
Protected Attributes inherited from Kernel
MatrixXd	Dbar
MatrixXd	Dbar2

Detailed Description

Stationary kernel with C^1 smooth realizations.

Member Function Documentation

compute_gram()

void compute_gram ( const std::vector< MatrixXd > &  dists2, const VectorXd &  theta_values, MatrixXd &  gram  )

overridevirtual

Compute a Gram matrix given a vector of squared distances and kernel hyperparameters.

Parameters

[in]	dists2	Vector of squared distance matrices.
[in]	theta_values	Vector of hyperparameters.
[in,out]	gram	Gram matrix.

Returns

Gram matrix.

Implements Kernel.

References Kernel::compute_Dbar().

compute_gram_derivs()

void compute_gram_derivs ( const MatrixXd &  gram, const std::vector< MatrixXd > &  dists2, const VectorXd &  theta_values, std::vector< MatrixXd > &  gram_derivs  )

overridevirtual

Compute the derivatives of the Gram matrix with respect to the kernel hyperparameters.

Parameters

[in]	gram	Gram Matrix
[in]	dists2	Vector of squared distance matrices.
[in]	theta_values	Vector of hyperparameters.
[in,out]	gram_derivs	Vector of Gram matrix derivatives.

Returns

Derivatives of the Gram matrix w.r.t. the hyperparameters.

Implements Kernel.

References Kernel::compute_Dbar().

compute_first_deriv_pred_gram()

MatrixXd compute_first_deriv_pred_gram ( const MatrixXd &  pred_gram, const std::vector< MatrixXd > &  mixed_dists, const VectorXd &  theta_values, const int  index  )

overridevirtual

Compute the first derivatve of the prediction matrix for a given component.

Parameters

[in]	pred_gram	Prediction Gram matrix - Rectangular matrix of kernel evaluations between the surrogate and prediction points.
[in]	mixed_dists	Component-wise signed distances between the prediction and build points.
[in]	theta_values	Vector of hyperparameters.
[in]	index	Specifies the component of the derivative.

Returns

first_deriv_pred_gram First derivative of the prediction Gram matrix for a given component.

Implements Kernel.

References dakota::surrogates::compute_cw_dists_squared(), Kernel::compute_Dbar(), and dakota::silence_unused_args().

compute_second_deriv_pred_gram()

MatrixXd compute_second_deriv_pred_gram ( const MatrixXd &  pred_gram, const std::vector< MatrixXd > &  mixed_dists, const VectorXd &  theta_values, const int  index_i, const int  index_j  )

overridevirtual

Compute the second derivatve of the prediction matrix for a pair of components.

Parameters

[in]	pred_gram	Prediction Gram matrix - Rectangular matrix of kernel evaluations between the surrogate and prediction points.
[in]	mixed_dists	Component-wise signed distances between the prediction and build points.
[in]	theta_values	Vector of hyperparameters.
[in]	index_i	Specifies the first component of the second derivative.
[in]	index_j	Specifies the second component of the second derivative.

Returns

second_deriv_pred_gram Second derivative of the prediction matrix for a pair of components.

Implements Kernel.

References dakota::silence_unused_args().

---

The documentation for this class was generated from the following files:

• SurrogatesGPKernels.hpp
• SurrogatesGPKernels.cpp