shark::DataDistribution< InputType > Class Template Reference

A DataDistribution defines an unsupervised learning problem. More...

#include <shark/Data/DataDistribution.h>

Public Member Functions
virtual	~DataDistribution ()
	Virtual destructor. More...
virtual void	draw (InputType &input) const =0
	Generates a single pair of input and label. More...
InputType	operator() ()
UnlabeledData< InputType >	generateDataset (std::size_t size, std::size_t maximumBatchSize) const
	Generates a data set with samples from from the distribution. More...
UnlabeledData< InputType >	generateDataset (std::size_t size) const
	Generates a data set with samples from from the distribution. More...

Detailed Description

template<class InputType>
class shark::DataDistribution< InputType >

A DataDistribution defines an unsupervised learning problem.

The unsupervised learning problem is defined by an explicit distribution (in contrast to a finite dataset). The only method we need is to draw a sample from the distribution.

Definition at line 57 of file DataDistribution.h.

Constructor & Destructor Documentation

~DataDistribution()

template<class InputType>

virtual shark::DataDistribution< InputType >::~DataDistribution ( )

inlinevirtual

Virtual destructor.

Definition at line 61 of file DataDistribution.h.

Member Function Documentation

draw()

template<class InputType>

virtual void shark::DataDistribution< InputType >::draw ( InputType &  input ) const

pure virtual

Generates a single pair of input and label.

Parameters

input

the generated input

Implemented in shark::ImagePatches, and shark::NormalDistributedPoints.

Referenced by shark::DataDistribution< RealVector >::generateDataset(), shark::LabeledDataDistribution< RealVector, unsigned int >::generateDataset(), shark::DataDistribution< RealVector >::operator()(), shark::LabeledDataDistribution< RealVector, unsigned int >::operator()(), shark::DataDistribution< RealVector >::~DataDistribution(), and shark::LabeledDataDistribution< RealVector, unsigned int >::~LabeledDataDistribution().

generateDataset() [1/2]

template<class InputType>

UnlabeledData<InputType> shark::DataDistribution< InputType >::generateDataset ( std::size_t  size, std::size_t  maximumBatchSize  ) const

inline

Generates a data set with samples from from the distribution.

Parameters

size	the number of samples in the dataset
maximumBatchSize	the maximum size of a batch

Definition at line 79 of file DataDistribution.h.

Referenced by shark::DataDistribution< RealVector >::generateDataset(), and shark::LabeledDataDistribution< RealVector, unsigned int >::generateDataset().

generateDataset() [2/2]

template<class InputType>

UnlabeledData<InputType> shark::DataDistribution< InputType >::generateDataset ( std::size_t  size ) const

inline

Generates a data set with samples from from the distribution.

Parameters

size	the number of samples in the dataset

Definition at line 93 of file DataDistribution.h.

operator()()

template<class InputType>

InputType shark::DataDistribution< InputType >::operator() ( )

inline

Definition at line 69 of file DataDistribution.h.

---

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

• include/shark/Data/DataDistribution.h