class gpr::LikGaussian

Overview

Gaussian noise with variance. More…

#include <LikGaussian.h>
 
class LikGaussian {
public:
    // construction
 
    LikGaussian();
    virtual ~LikGaussian();
 
    // methods
 
    double evaluateLogPrior();
    void evaluateLogPriorGradient(FieldMatrixd& lpg);
    void evaluateTrainingCovarianceMatrix(const Coord& x, FieldMatrixd& C);
    void setSigma2(const double value);
    double getSigma2() const;
    void setPriorParametersGaussian(const PriorBase& prior);
    void setPriorParametersSqrtt(const PriorBase& prior);
    void setParameters(const Eigen::VectorXd& w);
    Eigen::VectorXd combineParameters();
};

Detailed Documentation

Gaussian noise with variance.

Methods

double evaluateLogPrior()

breif Evaluate the log prior of covariance function parameters.

This function returns

\[log(p(th))\]

, where th collects the parameters. This function is needed when there are likelihood parameters.

void evaluateLogPriorGradient(FieldMatrixd& lpg)

Evaluate gradient of the log prior with respect to the parameters.

This function takes a Gaussian likelihood and returns

\[LPG = d log (p(th))/dth\]

, where th is the vector of parameters. This function is needed when there are likelihood parameters.

Parameters:

lpg

Result

void evaluateTrainingCovarianceMatrix(const Coord& x, FieldMatrixd& C)

Evaluate training covariance matrix of inputs.

Parameters:

x	Set of coordinates
C	Covariance matrix

void setSigma2(const double value)

Set sigma2 value.

double getSigma2() const

Get sigma2 value (the per-row noise variance added to the energy diagonal of the training kernel by evaluateTrainingCovarianceMatrix).

Eigen::VectorXd combineParameters()

Combine parameters of the gaussian likelihood into one Eigen vector.