BayesOpt
bayesopt::MCMCModel Class Reference

Posterior model of nonparametric processes/criteria based on MCMC samples. More...

#include <posterior_mcmc.hpp>

+ Inheritance diagram for bayesopt::MCMCModel:
+ Collaboration diagram for bayesopt::MCMCModel:

Public Types

typedef boost::ptr_vector< NonParametricProcessGPVect
 
typedef boost::ptr_vector< CriteriaCritVect
 

Public Member Functions

 MCMCModel (size_t dim, Parameters params, randEngine &eng)
 Constructor (Note: default constructor is private) More...
 
void updateHyperParameters ()
 
void fitSurrogateModel ()
 
void updateSurrogateModel ()
 
double evaluateCriteria (const vectord &query)
 
void updateCriteria (const vectord &query)
 
bool criteriaRequiresComparison ()
 
void setFirstCriterium ()
 
bool setNextCriterium (const vectord &prevResult)
 
std::string getBestCriteria (vectord &best)
 
ProbabilityDistributiongetPrediction (const vectord &query)
 
- Public Member Functions inherited from bayesopt::PosteriorModel
 PosteriorModel (size_t dim, Parameters params, randEngine &eng)
 Constructor. More...
 
virtual ~PosteriorModel ()
 Default destructor.
 
void setSamples (const matrixd &x, const vectord &y)
 
void setSamples (const matrixd &x)
 
void setSamples (const vectord &y)
 
void setSample (const vectord &x, double y)
 
void addSample (const vectord &x, double y)
 
double getValueAtMinimum ()
 
vectord getPointAtMinimum ()
 
void plotDataset (TLogLevel level)
 
const DatasetgetData ()
 

Private Member Functions

void setSurrogateModel (randEngine &eng)
 
void setCriteria (randEngine &eng)
 
 MCMCModel (MCMCModel &copy)
 

Private Attributes

size_t nParticles
 
GPVect mGP
 Pointer to surrogate model.
 
CritVect mCrit
 Metacriteria model.
 
boost::scoped_ptr< MCMCSamplerkSampler
 

Additional Inherited Members

- Static Public Member Functions inherited from bayesopt::PosteriorModel
static PosteriorModelcreate (size_t dim, Parameters params, randEngine &eng)
 
- Protected Attributes inherited from bayesopt::PosteriorModel
Parameters mParameters
 Configuration parameters.
 
size_t mDims
 Number of dimensions.
 
Dataset mData
 Dataset (x-> inputs, y-> labels/output)
 
MeanModel mMean
 

Detailed Description

Posterior model of nonparametric processes/criteria based on MCMC samples.

For computational reasons we store a copy of each conditional models with the corresponding particle generated by MCMC. That is to avoid costly operations like matrix inversions for every kernel parameter in a GP prediction. Thus, we assume that the number of particles is not very large.

Definition at line 47 of file posterior_mcmc.hpp.

Constructor & Destructor Documentation

◆ MCMCModel()

bayesopt::MCMCModel::MCMCModel ( size_t  dim,
Parameters  params,
randEngine &  eng 
)

Constructor (Note: default constructor is private)

Parameters
dimnumber of input dimensions
paramsconfiguration parameters (see parameters.hpp)
engrandom number generation engine (boost)

Definition at line 27 of file posterior_mcmc.cpp.

References mGP.


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