shark::IParameterizable Class Reference

Top level interface for everything that holds parameters. More...

#include <shark/Core/IParameterizable.h>

+ Inheritance diagram for shark::IParameterizable:

Public Member Functions

virtual ~IParameterizable ()
 
virtual RealVector parameterVector () const
 Return the parameter vector. More...
 
virtual void setParameterVector (RealVector const &newParameters)
 Set the parameter vector. More...
 
virtual std::size_t numberOfParameters () const
 Return the number of parameters. More...
 

Detailed Description

Top level interface for everything that holds parameters.

This interface is inherited by AbstractModel for unified access to the parameters of models, but also by objective functions and algorithms with hyper-parameters.

Definition at line 49 of file IParameterizable.h.

Constructor & Destructor Documentation

§ ~IParameterizable()

virtual shark::IParameterizable::~IParameterizable ( )
inlinevirtual

Definition at line 51 of file IParameterizable.h.

Member Function Documentation

§ numberOfParameters()

virtual std::size_t shark::IParameterizable::numberOfParameters ( ) const
inlinevirtual

Return the number of parameters.

Reimplemented in shark::KernelBudgetedSGDTrainer< InputType, CacheType >, shark::ConcatenatedModel< InputType, OutputType >, shark::AbstractLinearSvmTrainer< InputType >, shark::BinaryLayer, shark::BipolarLayer, shark::AbstractSvmTrainer< InputType, LabelType, Model, Trainer >, shark::AbstractSvmTrainer< InputType, unsigned int, MissingFeaturesKernelExpansion< InputType > >, shark::AbstractSvmTrainer< InputType, unsigned int >, shark::AbstractSvmTrainer< InputType, unsigned int, KernelClassifier< InputType >, AbstractWeightedTrainer< KernelClassifier< InputType > > >, shark::AbstractSvmTrainer< InputType, RealVector, KernelExpansion< InputType > >, shark::GaussianLayer, shark::TruncatedExponentialLayer, shark::KernelSGDTrainer< InputType, CacheType >, shark::KernelExpansion< InputType >, shark::KernelExpansion< RealVector >, shark::ModelKernel< InputType >, shark::WeightedSumKernel< InputType >, shark::Normalizer< DataType >, shark::FFNet< HiddenNeuron, OutputNeuron >, shark::ArgMaxConverter< Model >, shark::ArgMaxConverter< LinearModel< VectorType > >, shark::ArgMaxConverter< KernelExpansion< InputType > >, shark::RNNet, shark::CARTClassifier< LabelType >, shark::CARTClassifier< RealVector >, shark::PolynomialKernel< InputType >, shark::OneClassSvmTrainer< InputType, CacheType >, shark::LinearModel< InputType >, shark::LinearModel< VectorType >, shark::LassoRegression< InputVectorType >, shark::EpsilonSvmTrainer< InputType, CacheType >, shark::SoftNearestNeighborClassifier< InputType >, shark::ThresholdVectorConverter, shark::ConvolutionalRBM< VisibleLayerT, HiddenLayerT, RngT >, shark::ProductKernel< InputType >, shark::CMACMap, shark::NearestNeighborClassifier< InputType >, shark::ConvexCombination, shark::NearestNeighborRegression< InputType >, shark::RBM< VisibleLayerT, HiddenLayerT, RngT >, shark::HierarchicalClustering< InputT >, shark::TiedAutoencoder< HiddenNeuron, OutputNeuron >, shark::Autoencoder< HiddenNeuron, OutputNeuron >, shark::ARDKernelUnconstrained< InputType >, shark::OneVersusOneClassifier< InputType >, shark::NormalizedKernel< InputType >, shark::GaussianRbfKernel< InputType >, shark::OnlineRNNet, shark::Softmax, shark::GaussianNoiseModel, shark::RBFLayer, shark::ImpulseNoiseModel, shark::LDA, shark::ThresholdConverter, shark::DiscreteKernel, shark::MonomialKernel< InputType >, shark::Centroids, shark::LinearRegression, shark::ScaledKernel< InputType >, shark::SigmoidModel, shark::ClusteringModel< InputT, OutputT >, shark::ClusteringModel< InputT, RealVector >, shark::ClusteringModel< InputT, unsigned int >, and shark::LinearNorm.

Definition at line 66 of file IParameterizable.h.

References parameterVector().

Referenced by shark::ProductKernel< InputType >::addKernel(), shark::calculateKernelMatrixParameterDerivative(), shark::RadiusMarginQuotient< InputType, CacheType >::eval(), shark::NegativeGaussianProcessEvidence< InputType, OutputType, LabelType >::eval(), shark::RadiusMarginQuotient< InputType, CacheType >::evalDerivative(), shark::NegativeGaussianProcessEvidence< InputType, OutputType, LabelType >::evalDerivative(), shark::KernelTargetAlignment< InputType, LabelType >::evalDerivative(), shark::initRandomNormal(), shark::initRandomUniform(), shark::ClusteringModel< InputT, unsigned int >::numberOfParameters(), shark::ScaledKernel< InputType >::numberOfParameters(), shark::NormalizedKernel< InputType >::numberOfParameters(), shark::OneClassSvmTrainer< InputType, CacheType >::numberOfParameters(), shark::KernelSGDTrainer< InputType, CacheType >::numberOfParameters(), shark::KernelBudgetedSGDTrainer< InputType, CacheType >::numberOfParameters(), shark::LooErrorCSvm< InputType, CacheType >::numberOfVariables(), shark::NegativeLogLikelihood::numberOfVariables(), shark::RadiusMarginQuotient< InputType, CacheType >::numberOfVariables(), shark::NegativeGaussianProcessEvidence< InputType, OutputType, LabelType >::numberOfVariables(), shark::LooError< ModelTypeT, LabelType >::numberOfVariables(), shark::CrossValidationError< ModelTypeT, LabelTypeT >::numberOfVariables(), shark::KernelTargetAlignment< InputType, LabelType >::numberOfVariables(), shark::OneClassSvmTrainer< InputType, CacheType >::parameterVector(), shark::KernelSGDTrainer< InputType, CacheType >::parameterVector(), shark::KernelBudgetedSGDTrainer< InputType, CacheType >::parameterVector(), shark::RFTrainer::setParameterVector(), shark::OneClassSvmTrainer< InputType, CacheType >::setParameterVector(), shark::KernelSGDTrainer< InputType, CacheType >::setParameterVector(), shark::KernelBudgetedSGDTrainer< InputType, CacheType >::setParameterVector(), shark::NegativeGaussianProcessEvidence< InputType, OutputType, LabelType >::setThreshold(), and shark::CSvmTrainer< InputType, CacheType >::train().

§ parameterVector()

virtual RealVector shark::IParameterizable::parameterVector ( ) const
inlinevirtual

Return the parameter vector.

Reimplemented in shark::KernelBudgetedSGDTrainer< InputType, CacheType >, shark::ConcatenatedModel< InputType, OutputType >, shark::AbstractLinearSvmTrainer< InputType >, shark::BinaryLayer, shark::BipolarLayer, shark::GaussianLayer, shark::TruncatedExponentialLayer, shark::AbstractSvmTrainer< InputType, LabelType, Model, Trainer >, shark::AbstractSvmTrainer< InputType, unsigned int, MissingFeaturesKernelExpansion< InputType > >, shark::AbstractSvmTrainer< InputType, unsigned int >, shark::AbstractSvmTrainer< InputType, unsigned int, KernelClassifier< InputType >, AbstractWeightedTrainer< KernelClassifier< InputType > > >, shark::AbstractSvmTrainer< InputType, RealVector, KernelExpansion< InputType > >, shark::KernelSGDTrainer< InputType, CacheType >, shark::ModelKernel< InputType >, shark::FFNet< HiddenNeuron, OutputNeuron >, shark::KernelExpansion< InputType >, shark::KernelExpansion< RealVector >, shark::NBClassifier< InputType, OutputType >, shark::ArgMaxConverter< Model >, shark::ArgMaxConverter< LinearModel< VectorType > >, shark::ArgMaxConverter< KernelExpansion< InputType > >, shark::CARTClassifier< LabelType >, shark::CARTClassifier< RealVector >, shark::Normalizer< DataType >, shark::RNNet, shark::WeightedSumKernel< InputType >, shark::ConvolutionalRBM< VisibleLayerT, HiddenLayerT, RngT >, shark::RBM< VisibleLayerT, HiddenLayerT, RngT >, shark::ThresholdVectorConverter, shark::OneClassSvmTrainer< InputType, CacheType >, shark::TiedAutoencoder< HiddenNeuron, OutputNeuron >, shark::LinearModel< InputType >, shark::LinearModel< VectorType >, shark::Autoencoder< HiddenNeuron, OutputNeuron >, shark::LassoRegression< InputVectorType >, shark::SoftNearestNeighborClassifier< InputType >, shark::EpsilonSvmTrainer< InputType, CacheType >, shark::CMACMap, shark::ProductKernel< InputType >, shark::RFTrainer, shark::HierarchicalClustering< InputT >, shark::NearestNeighborClassifier< InputType >, shark::NearestNeighborRegression< InputType >, shark::ConvexCombination, shark::ARDKernelUnconstrained< InputType >, shark::PolynomialKernel< InputType >, shark::NormalizedKernel< InputType >, shark::OnlineRNNet, shark::RBFLayer, shark::ThresholdConverter, shark::GaussianNoiseModel, shark::Softmax, shark::ImpulseNoiseModel, shark::MonomialKernel< InputType >, shark::LDA, shark::Centroids, shark::OneVersusOneClassifier< InputType >, shark::DiscreteKernel, shark::GaussianRbfKernel< InputType >, shark::LinearRegression, shark::ScaledKernel< InputType >, shark::SigmoidModel, shark::MeanModel< ModelType >, shark::MeanModel< CARTClassifier< RealVector > >, shark::ClusteringModel< InputT, OutputT >, shark::LinearKernel< InputType >, shark::ClusteringModel< InputT, RealVector >, shark::ClusteringModel< InputT, unsigned int >, and shark::LinearNorm.

Definition at line 54 of file IParameterizable.h.

Referenced by numberOfParameters(), shark::ClusteringModel< InputT, unsigned int >::parameterVector(), shark::ScaledKernel< InputType >::parameterVector(), shark::NormalizedKernel< InputType >::parameterVector(), shark::OneClassSvmTrainer< InputType, CacheType >::parameterVector(), shark::NegativeLogLikelihood::proposeStartingPoint(), shark::AbstractMetric< InputType >::write(), and shark::AbstractModel< InputT, unsigned int >::write().

§ setParameterVector()

virtual void shark::IParameterizable::setParameterVector ( RealVector const &  newParameters)
inlinevirtual

Set the parameter vector.

Reimplemented in shark::KernelBudgetedSGDTrainer< InputType, CacheType >, shark::ConcatenatedModel< InputType, OutputType >, shark::AbstractLinearSvmTrainer< InputType >, shark::BinaryLayer, shark::BipolarLayer, shark::AbstractSvmTrainer< InputType, LabelType, Model, Trainer >, shark::AbstractSvmTrainer< InputType, unsigned int, MissingFeaturesKernelExpansion< InputType > >, shark::AbstractSvmTrainer< InputType, unsigned int >, shark::AbstractSvmTrainer< InputType, unsigned int, KernelClassifier< InputType >, AbstractWeightedTrainer< KernelClassifier< InputType > > >, shark::AbstractSvmTrainer< InputType, RealVector, KernelExpansion< InputType > >, shark::GaussianLayer, shark::TruncatedExponentialLayer, shark::KernelSGDTrainer< InputType, CacheType >, shark::ModelKernel< InputType >, shark::FFNet< HiddenNeuron, OutputNeuron >, shark::KernelExpansion< InputType >, shark::KernelExpansion< RealVector >, shark::NBClassifier< InputType, OutputType >, shark::ArgMaxConverter< Model >, shark::ArgMaxConverter< LinearModel< VectorType > >, shark::ArgMaxConverter< KernelExpansion< InputType > >, shark::Normalizer< DataType >, shark::WeightedSumKernel< InputType >, shark::CARTClassifier< LabelType >, shark::CARTClassifier< RealVector >, shark::RNNet, shark::ConvolutionalRBM< VisibleLayerT, HiddenLayerT, RngT >, shark::RBM< VisibleLayerT, HiddenLayerT, RngT >, shark::OneClassSvmTrainer< InputType, CacheType >, shark::LinearModel< InputType >, shark::LinearModel< VectorType >, shark::TiedAutoencoder< HiddenNeuron, OutputNeuron >, shark::Autoencoder< HiddenNeuron, OutputNeuron >, shark::ThresholdVectorConverter, shark::LassoRegression< InputVectorType >, shark::EpsilonSvmTrainer< InputType, CacheType >, shark::SoftNearestNeighborClassifier< InputType >, shark::PolynomialKernel< InputType >, shark::RFTrainer, shark::ProductKernel< InputType >, shark::CMACMap, shark::HierarchicalClustering< InputT >, shark::NearestNeighborClassifier< InputType >, shark::NearestNeighborRegression< InputType >, shark::ConvexCombination, shark::ARDKernelUnconstrained< InputType >, shark::NormalizedKernel< InputType >, shark::OneVersusOneClassifier< InputType >, shark::OnlineRNNet, shark::RBFLayer, shark::GaussianNoiseModel, shark::GaussianRbfKernel< InputType >, shark::ThresholdConverter, shark::Softmax, shark::LDA, shark::ImpulseNoiseModel, shark::DiscreteKernel, shark::MonomialKernel< InputType >, shark::Centroids, shark::LinearRegression, shark::ScaledKernel< InputType >, shark::MeanModel< ModelType >, shark::SigmoidModel, shark::MeanModel< CARTClassifier< RealVector > >, shark::ClusteringModel< InputT, OutputT >, shark::ClusteringModel< InputT, RealVector >, shark::ClusteringModel< InputT, unsigned int >, shark::LinearKernel< InputType >, and shark::LinearNorm.

Definition at line 60 of file IParameterizable.h.

References SHARK_ASSERT.

Referenced by shark::NegativeLogLikelihood::eval(), shark::RadiusMarginQuotient< InputType, CacheType >::eval(), shark::NegativeGaussianProcessEvidence< InputType, OutputType, LabelType >::eval(), shark::CrossValidationError< ModelTypeT, LabelTypeT >::eval(), shark::SvmLogisticInterpretation< InputType >::eval(), shark::LooError< ModelTypeT, LabelType >::eval(), shark::KernelTargetAlignment< InputType, LabelType >::eval(), shark::NegativeLogLikelihood::evalDerivative(), shark::RadiusMarginQuotient< InputType, CacheType >::evalDerivative(), shark::NegativeGaussianProcessEvidence< InputType, OutputType, LabelType >::evalDerivative(), shark::KernelTargetAlignment< InputType, LabelType >::evalDerivative(), shark::SvmLogisticInterpretation< InputType >::evalDerivative(), shark::initRandomNormal(), shark::initRandomUniform(), shark::AbstractMetric< InputType >::read(), shark::AbstractModel< InputT, unsigned int >::read(), shark::ClusteringModel< InputT, unsigned int >::setParameterVector(), shark::ScaledKernel< InputType >::setParameterVector(), shark::NormalizedKernel< InputType >::setParameterVector(), and shark::OneClassSvmTrainer< InputType, CacheType >::setParameterVector().


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