Hinge-loss for large margin regression. More...

#include <shark/ObjectiveFunctions/Loss/EpsilonHingeLoss.h>

Inheritance diagram for shark::EpsilonHingeLoss:

Public Member Functions
	EpsilonHingeLoss (double epsilon)
	constructor More...

std::string	name () const
	Returns class name "HingeLoss". More...

double	eval (BatchLabelType const &labels, BatchOutputType const &predictions) const
	calculates the sum of all More...

double	evalDerivative (BatchLabelType const &labels, BatchOutputType const &predictions, BatchOutputType &gradient) const
	evaluate the loss and the derivative w.r.t. the prediction More...

Public Member Functions inherited from shark::AbstractLoss< RealVector, RealVector >
	AbstractLoss ()

virtual double	eval (LabelType const &target, OutputType const &prediction) const
	evaluate the loss for a target and a prediction More...

double	eval (Data< LabelType > const &targets, Data< OutputType > const &predictions) const

virtual double	evalDerivative (LabelType const &target, OutputType const &prediction, OutputType &gradient) const
	evaluate the loss and its derivative for a target and a prediction More...

virtual double	evalDerivative (LabelType const &target, OutputType const &prediction, OutputType &gradient, MatrixType &hessian) const
	evaluate the loss and its first and second derivative for a target and a prediction More...

double	operator() (LabelType const &target, OutputType const &prediction) const
	evaluate the loss for a target and a prediction More...

double	operator() (BatchLabelType const &target, BatchOutputType const &prediction) const

Public Member Functions inherited from shark::AbstractCost< RealVector, RealVector >
virtual	~AbstractCost ()

const Features &	features () const

virtual void	updateFeatures ()

bool	hasFirstDerivative () const
	returns true when the first parameter derivative is implemented More...

bool	isLossFunction () const
	returns true when the cost function is in fact a loss function More...

double	operator() (Data< LabelType > const &targets, Data< OutputType > const &predictions) const

Public Member Functions inherited from shark::INameable
virtual	~INameable ()

Additional Inherited Members
Public Types inherited from shark::AbstractLoss< RealVector, RealVector >
typedef RealVector	OutputType

typedef RealVector	LabelType

typedef VectorMatrixTraits< OutputType >::DenseMatrixType	MatrixType

typedef Batch< OutputType >::type	BatchOutputType

typedef Batch< LabelType >::type	BatchLabelType

Public Types inherited from shark::AbstractCost< RealVector, RealVector >
enum	Feature
	list of features a cost function can have More...

typedef RealVector	OutputType

typedef RealVector	LabelType

typedef Batch< OutputType >::type	BatchOutputType

typedef Batch< LabelType >::type	BatchLabelType

typedef TypedFlags< Feature >	Features

typedef TypedFeatureNotAvailableException< Feature >	FeatureNotAvailableException

Protected Attributes inherited from shark::AbstractCost< RealVector, RealVector >
Features	m_features

Detailed Description

Hinge-loss for large margin regression.

The loss is defined as \(L_i = \sum_j^N \max\{0.0, |f_j(x_i)-y_{i,j}|-\epsilon\} \) where \( y_i =(y_{i,1},\dots,y_{i_N} \) is the label of dimension N and \( f_j(x_i) \) is the j-th output of the prediction of the model for the ith input. The loss introduces the concept of a margin to regression, that is, points are not punished when they are sufficiently close to the function. Points which are outside of the margin are linearly punished, that is the loss is outlier resistant.

Epsilon describes the size of the margin.

The hinge-loss is not differentiable at the points y_{i,j}+epsilon and y_{i,j}-epsilon.

Definition at line 51 of file EpsilonHingeLoss.h.

Constructor & Destructor Documentation

§ EpsilonHingeLoss()

shark::EpsilonHingeLoss::EpsilonHingeLoss ( double epsilon )

inline

constructor

Definition at line 55 of file EpsilonHingeLoss.h.

References shark::AbstractCost< RealVector, RealVector >::m_features.

Member Function Documentation

§ eval()

double shark::EpsilonHingeLoss::eval	(	BatchLabelType const &	labels,
		BatchOutputType const &	predictions
	)		const

inlinevirtual

calculates the sum of all

Implements shark::AbstractLoss< RealVector, RealVector >.

Definition at line 65 of file EpsilonHingeLoss.h.

References shark::blas::max(), SIZE_CHECK, and shark::blas::sum().

§ evalDerivative()

double shark::EpsilonHingeLoss::evalDerivative	(	BatchLabelType const &	target,
		BatchOutputType const &	prediction,
		BatchOutputType &	gradient
	)		const

inlinevirtual

evaluate the loss and the derivative w.r.t. the prediction

: The default implementations throws an exception. If you overwrite this method, don't forget to set the flag HAS_FIRST_DERIVATIVE.

Parameters

target	target value
prediction	prediction, typically made by a model
gradient	the gradient of the loss function with respect to the prediction

Reimplemented from shark::AbstractLoss< RealVector, RealVector >.

Definition at line 71 of file EpsilonHingeLoss.h.

References shark::blas::max(), and SIZE_CHECK.

§ name()

std::string shark::EpsilonHingeLoss::name ( ) const

inlinevirtual

Returns class name "HingeLoss".

Reimplemented from shark::INameable.

Definition at line 60 of file EpsilonHingeLoss.h.

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

include/shark/ObjectiveFunctions/Loss/EpsilonHingeLoss.h

Public Member Functions

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

§ EpsilonHingeLoss()

Member Function Documentation

§ eval()

§ evalDerivative()

§ name()