bayesopt::NonParametricProcess Class Reference

Abstract class to implement Bayesian regressors. More...

#include <nonparametricprocess.hpp>

Inheritance diagram for bayesopt::NonParametricProcess:

Collaboration diagram for bayesopt::NonParametricProcess:

Public Member Functions
	NonParametricProcess (size_t dim, Parameters parameters, const Dataset &data, MeanModel &mean, randEngine &eng)
virtual ProbabilityDistribution *	prediction (const vectord &query)=0
	Function that returns the prediction of the GP for a query point in the hypercube [0,1]. More...
virtual void	fitSurrogateModel ()=0
	Computes the initial surrogate model and updates the kernel parameters estimation. More...
virtual void	updateSurrogateModel ()=0
	Sequential update of the surrogate model by adding a new row to the Kernel matrix, more precisely, to its Cholesky decomposition. More...
double	getValueAtMinimum ()
const Dataset *	getData ()
double	getSignalVariance ()
virtual size_t	nHyperParameters ()=0
virtual vectord	getHyperParameters ()=0
virtual void	setHyperParameters (const vectord &theta)=0
Public Member Functions inherited from bayesopt::RBOptimizable
virtual double	evaluate (const vectord &query)=0

Static Public Member Functions
static NonParametricProcess *	create (size_t dim, Parameters parameters, const Dataset &data, MeanModel &mean, randEngine &eng)
	Factory model generator for surrogate models. More...

Protected Attributes
const Dataset &	mData
double	mSigma
	Signal variance.
size_t	dim_
MeanModel &	mMean

Detailed Description

Abstract class to implement Bayesian regressors.

Definition at line 43 of file nonparametricprocess.hpp.

Member Function Documentation

create()

NonParametricProcess * bayesopt::NonParametricProcess::create ( size_t  dim, Parameters  parameters, const Dataset &  data, MeanModel &  mean, randEngine &  eng  )

static

Factory model generator for surrogate models.

Parameters

parameters

(process name, noise, priors, etc.)

Returns

pointer to the corresponding derivate class (surrogate model)

Definition at line 49 of file nonparametricprocess.cpp.

References bayesopt::Parameters::surr_name.

Referenced by bayesopt::PosteriorFixed::~PosteriorFixed().

fitSurrogateModel()

virtual void bayesopt::NonParametricProcess::fitSurrogateModel ( )

pure virtual

Computes the initial surrogate model and updates the kernel parameters estimation.

This function requires to recompute all covariance matrixes, inverses, etc. Use it with precaution.

Implemented in bayesopt::KernelRegressor.

prediction()

virtual ProbabilityDistribution* bayesopt::NonParametricProcess::prediction ( const vectord &  query )

pure virtual

Function that returns the prediction of the GP for a query point in the hypercube [0,1].

Parameters

query

in the hypercube [0,1] to evaluate the Gaussian process

Returns

pointer to the probability distribution.

Implemented in bayesopt::StudentTProcessNIG, bayesopt::ConditionalBayesProcess, bayesopt::GaussianProcessML, bayesopt::GaussianProcessNormal, bayesopt::StudentTProcessJeffreys, and bayesopt::GaussianProcess.

updateSurrogateModel()

virtual void bayesopt::NonParametricProcess::updateSurrogateModel ( )

pure virtual

Sequential update of the surrogate model by adding a new row to the Kernel matrix, more precisely, to its Cholesky decomposition.

It assumes that the kernel hyperparemeters do not change.

Implemented in bayesopt::KernelRegressor.

---

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

• nonparametricprocess.hpp
• nonparametricprocess.cpp