Wrapper for the dataset data. More...

#include <Data.hpp>

Public Member Functions
const std::string &	getType () const
	Returns the type of the dataset. More...

void	setType (const std::string &type)

	Data (const Data< T > &other)

	Data (const std::string &dataset, bool atEnd)

	Data (const char dataset, const char datasetType, const char pos_class="1", const char neg_class="-1")
	Data constructor to load a dataset from a file. More...

	Data (size_t size, size_t dim, T val=T())
	Constructor for instantiation with an initial size and dimension. More...

	Data (const char *dataset)
	Constructor for empty data. More...

bool	isClassification () const
	Inform if the dataset is used for classification. More...

std::string	name () const
	The name of the dataset is defined as the name ofthe file where it were loaded from. More...

size_t	size () const
	Returns the size of the dataset. More...

size_t	dim () const
	Returns the dimension of the dataset. More...

std::vector< std::shared_ptr< Point< T > > >	points ()
	Returns a shared pointer to the vector of Points of the sample. More...

std::vector< std::shared_ptr< Point< T > > >	points () const
	Returns a shared pointer to the vector of Points of the sample. More...

const std::vector< int >	classes () const
	Returns a vector containing the numeric values of the classes. More...

PointPointer< T >	point (int index) const
	Returns a shared pointer to the point with the given index. More...

Point< T >	getFeature (int index) const
	Get the values of a feature from all points. More...

Point< double >	getLabels () const

Point< double >	labels () const

std::vector< size_t >	classesDistribution () const
	Returns a vector containing the frequency of the classes. Only valid for classification datasets. More...

std::vector< std::string >	classesNames () const
	Returns a vector containing the name of the classes. Only valid for classification datasets. More...

std::vector< int >	getFeaturesNames () const
	Returns the features names. More...

mltk::Point< int >	featuresNames () const

std::vector< int >	getIndex () const
	Returns the vector of indexes. More...

double	getTime_mult () const
	Return the time multiplier. More...

void	setName (const std::string &name)
	Set the name of the dataset. More...

void	setClassesNames (const std::vector< std::string > class_names)
	Set classes names. More...

void	setClasses (const std::vector< int > &classes)
	Set the classes to use in the dataset. More...

void	setClassesAtEnd (bool atEnd)
	Inform if the classes of the dataset to be loaded are at begining or end of the attributes. More...

void	setPoint (int index, std::shared_ptr< Point< T > > p)
	setPoint Set the point in a position of the data. More...

void	setFeaturesNames (const std::vector< int > &fnames)
	setFeaturesNames Set the name of the features of the data. More...

void	setIndex (std::vector< int > index)
	Set the index vector for the data. More...

void	setDim (size_t dim)
	setDim Set the dimension of the points. More...

void	computeClassesDistribution ()
	Compute the frequency of each class in the dataset. More...

void	shuffle (const size_t &seed=42)
	Shuffle the data with a given seed. More...

bool	load (const std::string &file)
	Load a dataset from a file. More...

bool	load (const std::string &file, bool _atEnd)
	Load a dataset from a file. More...

void	write (const std::string &fname, std::string ext)
	write Write the data to a file with the given extention. More...

bool	isEmpty () const
	Returns if there's a dataset loaded. More...

bool	isNormalized () const
	Returns if the dataset is normalized. More...

void	clear ()
	clear Clear the data. More...

void	classesCopy (const Data< T > &_data, std::vector< int > &classes)
	Makes a deep copy from another data object. More...

mltk::Data< T >	copy () const
	Returns a copy of itself. More...

void	copy (const Data< T > &_data)
	Makes a deep copy from another data object. More...

void	copyZero (const Data< T > &other)
	Returns a copy of the data with zero points. More...

std::vector< Data< T > >	splitByClasses (bool keepIndex=false)
	Split the dataset by its labels. Only valid for classification datasets. More...

std::vector< Data< T > >	splitSample (const std::size_t &split_size, bool stratified=true, bool keepIndex=false, size_t seed=0)
	Split the data by a given size. More...

Data< T >	selectFeatures (std::vector< size_t > feats, int size=-1)
	Returns a Data object with selected features. More...

Data< T >	sampling (const size_t &samp_size, bool with_replacement=true, const int &seed=0)
	Sample the dataset with the given size. More...

void	apply (std::function< void(mltk::PointPointer< T > point)> f)
	Apply a function to all points on the dataset. More...

void	join (const Data< T > &data)
	Merge one dataset with another. More...

bool	insertPoint (const Data< T > &samples, int _index, bool keepIndex=false)
	Insert a point to the data from another sample. More...

bool	insertPoint (std::shared_ptr< Point< T > > p, bool keepIndex=false)
	Insert a point to the end of points vector. More...

bool	insertPoint (Point< T > p, bool keepIndex=false)
	Insert a point to the end of points vector. More...

std::vector< bool >	removePoints (std::vector< int > ids)
	Remove several points from the sample. More...

bool	removePoint (int pid)
	Remove a point from the data. More...

Data< T >	insertFeatures (std::vector< int > ins_feat, bool keepIndex=false)
	insertFeatures Returns Data object with only features in array. More...

Data< T >	removeFeatures (std::vector< int > feats, int fsize) const
	Remove several features from the sample. More...

bool	removeFeatures (std::vector< int > feats)
	Remove features from the dataset. More...

bool	updatePointValue (const size_t &idx, double value)
	Updates a Point value. More...

void	changeXVector (std::vector< int > index)
	Change the x vector of a sample. More...

void	normalize (double p=2)
	normalize Normalize the dataset using a Lp-norm. More...

void	resetIndex ()
	Reset the index vector. More...

SampleIterator< T >	begin ()

SampleIterator< T >	end ()

ConstSampleIterator< T >	begin () const

ConstSampleIterator< T >	end () const

std::shared_ptr< Point< T > >	operator[] (size_t i) const

std::shared_ptr< Point< T > > &	operator[] (size_t i)

Point< T >	operator() (size_t i) const

Point< T > &	operator() (size_t i)

Data< T > &	operator= (const Data< T > &)

bool	operator== (const Data< T > &rhs) const

bool	operator!= (const Data< T > &rhs) const

Friends
template<typename U >
std::ostream &	operator<< (std::ostream &output, const Data< U > &data)

Detailed Description

template<typename T = double>
class mltk::Data< T >

Wrapper for the dataset data.

Constructor & Destructor Documentation

◆ Data() [1/4]

template<typename T >

mltk::Data< T >::Data	(	const std::string &	dataset,
		bool	atEnd
	)

Parameters

dataset	Dataset path.
atEnd	Indicates if labels are at end of the columns of the dataset.

◆ Data() [2/4]

template<typename T >

mltk::Data< T >::Data	(	const char *	dataset,
		const char *	datasetType,
		const char *	pos_class = `"1"`,
		const char *	neg_class = `"-1"`
	)

explicit

Data constructor to load a dataset from a file.

Parameters

dataset	Path to the dataset to be loaded.
pos_class	String representing the positive class on the dataset.
neg_class	String representing the negative class on the dataset.

◆ Data() [3/4]

template<typename T >

mltk::Data< T >::Data	(	size_t	size,
		size_t	dim,
		T	val = `T()`
	)

Constructor for instantiation with an initial size and dimension.

Parameters

size	The initial size of the data.
dim	The initial dimension of the data.
val	The default value for the initialization, if not given, it'll call the default constructor of T.

◆ Data() [4/4]

template<typename T >

mltk::Data< T >::Data ( const char * dataset )

explicit

Constructor for empty data.

Parameters

pos_class	String representing the positive class on the dataset.
neg_class	String representing the negative class on the dataset.

Member Function Documentation

◆ apply()

template<typename T >

void mltk::Data< T >::apply ( std::function< void(mltk::PointPointer< T > point)> f )

Apply a function to all points on the dataset.

Parameters

f	Function to be applied to the data, must receive a PointPointer.

◆ changeXVector()

template<typename T >

void mltk::Data< T >::changeXVector ( std::vector< int > index )

Change the x vector of a sample.

Parameters

index Indexes of the change to be made.

Returns: void

◆ classes()

template<typename T >

const std::vector< int > mltk::Data< T >::classes

Returns a vector containing the numeric values of the classes.

Returns: std::vector<int>

◆ classesCopy()

template<typename T >

void mltk::Data< T >::classesCopy	(	const Data< T > &	_data,
		std::vector< int > &	classes
	)

Makes a deep copy from another data object.

Returns: Data

◆ classesDistribution()

template<typename T >

std::vector< size_t > mltk::Data< T >::classesDistribution

Returns a vector containing the frequency of the classes. Only valid for classification datasets.

Returns: std::vector<double> containing the distribution of the classes.

◆ classesNames()

template<typename T >

std::vector< std::string > mltk::Data< T >::classesNames

Returns a vector containing the name of the classes. Only valid for classification datasets.

Returns: std::vector<std::string> containing the names of the classes.

◆ clear()

template<typename T >

void mltk::Data< T >::clear

clear Clear the data.

◆ computeClassesDistribution()

template<typename T >

void mltk::Data< T >::computeClassesDistribution

Compute the frequency of each class in the dataset.

◆ copy() [1/2]

template<typename T >

Data< T > mltk::Data< T >::copy

Returns a copy of itself.

Returns: Data

◆ copy() [2/2]

template<typename T >

void mltk::Data< T >::copy ( const Data< T > & _data )

Makes a deep copy from another data object.

Returns: Data

◆ copyZero()

template<typename T >

void mltk::Data< T >::copyZero ( const Data< T > & other )

Returns a copy of the data with zero points.

Returns: Data

◆ dim()

template<typename T = double>

size_t mltk::Data< T >::dim ( ) const

inline

Returns the dimension of the dataset.

Returns: int

◆ getFeature()

template<typename T >

Point< T > mltk::Data< T >::getFeature ( int index ) const

Get the values of a feature from all points.

Parameters

index Index of the feature on the dataset.

Returns: Point containing feature values.

◆ getFeaturesNames()

template<typename T >

std::vector< int > mltk::Data< T >::getFeaturesNames

Returns the features names.

Returns: std::vector<int>

◆ getIndex()

template<typename T >

std::vector< int > mltk::Data< T >::getIndex

Returns the vector of indexes.

Returns: std::vector<int>

◆ getTime_mult()

template<typename T >

double mltk::Data< T >::getTime_mult

Return the time multiplier.

Returns: double

◆ getType()

template<typename T >

const std::string & mltk::Data< T >::getType

Returns the type of the dataset.

Returns: std::string

◆ insertFeatures()

template<typename T >

mltk::Data< T > mltk::Data< T >::insertFeatures	(	std::vector< int >	ins_feat,
		bool	keepIndex = `false`
	)

insertFeatures Returns Data object with only features in array.

Parameters

ins_feat Array with features that will be in the Data object.

Returns: Data If the object is empty something wrong happened.

◆ insertPoint() [1/3]

template<typename T >

bool mltk::Data< T >::insertPoint	(	const Data< T > &	samples,
		int	_index,
		bool	keepIndex = `false`
	)

Insert a point to the data from another sample.

Parameters

sample	Sample with the point to be added.
_index	Index of the point to be added.

Returns: bool

◆ insertPoint() [2/3]

template<typename T >

bool mltk::Data< T >::insertPoint	(	Point< T >	p,
		bool	keepIndex = `false`
	)

Insert a point to the end of points vector.

Parameters

p	Point to be inserted.

Returns: bool

◆ insertPoint() [3/3]

template<typename T >

bool mltk::Data< T >::insertPoint	(	std::shared_ptr< Point< T > >	p,
		bool	keepIndex = `false`
	)

Insert a point to the end of points vector.

Parameters

p	Point to be inserted.

Returns: bool

◆ isClassification()

template<typename T = double>

bool mltk::Data< T >::isClassification ( ) const

inline

Inform if the dataset is used for classification.

◆ isEmpty()

template<typename T >

bool mltk::Data< T >::isEmpty

Returns if there's a dataset loaded.

Returns: bool

◆ isNormalized()

template<typename T = double>

bool mltk::Data< T >::isNormalized ( ) const

inline

Returns if the dataset is normalized.

Returns: bool

◆ join()

template<typename T >

void mltk::Data< T >::join ( const Data< T > & data )

Merge one dataset with another.

Parameters

data	Dataset to be joined.

Returns: bool

◆ load() [1/2]

template<typename T >

bool mltk::Data< T >::load ( const std::string & file )

Load a dataset from a file.

Parameters

file	Path to dataset file.

Returns: bool

◆ load() [2/2]

template<typename T >

bool mltk::Data< T >::load	(	const std::string &	file,
		bool	_atEnd
	)

Load a dataset from a file.

Parameters

file	Path to dataset file.

Returns: bool

◆ name()

template<typename T = double>

std::string mltk::Data< T >::name ( ) const

inline

The name of the dataset is defined as the name ofthe file where it were loaded from.

Returns: Name of the dataset.

◆ normalize()

template<typename T >

void mltk::Data< T >::normalize ( double p = 2 )

normalize Normalize the dataset using a Lp-norm.

Parameters

p	Norm to be utilized.

◆ point()

template<typename T >

PointPointer< T > mltk::Data< T >::point ( int index ) const

Returns a shared pointer to the point with the given index.

Parameters

index Position of a point in the points array.

Returns: std::vector<Point< T > >

◆ points() [1/2]

template<typename T >

std::vector< std::shared_ptr< Point< T > > > mltk::Data< T >::points

Returns a shared pointer to the vector of Points of the sample.

Returns: std::vector<std::shared_ptr<Point< T > > >

◆ points() [2/2]

template<typename T >

std::vector< std::shared_ptr< Point< T > > > mltk::Data< T >::points

Returns a shared pointer to the vector of Points of the sample.

Returns: std::vector<std::shared_ptr<Point< T > > >

◆ removeFeatures() [1/2]

template<typename T >

bool mltk::Data< T >::removeFeatures ( std::vector< int > feats )

Remove features from the dataset.

Parameters

feats Features to be removed.

Returns: Success of feature removal.

◆ removeFeatures() [2/2]

template<typename T >

Data< T > mltk::Data< T >::removeFeatures	(	std::vector< int >	feats,
		int	fsize
	)		const

Remove several features from the sample.

Parameters

feats Names of the features to be removed (must be sorted).

Returns: boolean informing if all features were succesfully removed.

◆ removePoint()

template<typename T >

bool mltk::Data< T >::removePoint ( int pid )

Remove a point from the data.

Parameters

pid	Index of the point to be removed.

Returns: bool

◆ removePoints()

template<typename T >

std::vector< bool > mltk::Data< T >::removePoints ( std::vector< int > ids )

Remove several points from the sample.

Parameters

ids	Ids of the points to be removed (must be sorted).

Returns: booleans informing which points were removed succesfully.

◆ resetIndex()

template<typename T >

void mltk::Data< T >::resetIndex

Reset the index vector.

◆ sampling()

template<typename T >

Data< T > mltk::Data< T >::sampling	(	const size_t &	samp_size,
		bool	with_replacement = `true`,
		const int &	seed = `0`
	)

Sample the dataset with the given size.

Parameters

samp_size	Sampling size.
with_replacement	Tells if sampling must be made with replacement.
seed	Random generator seed.

Returns: Data object with sampled data.

◆ selectFeatures()

template<typename T >

Data< T > mltk::Data< T >::selectFeatures	(	std::vector< size_t >	feats,
		int	size = `-1`
	)

Returns a Data object with selected features.

Parameters

feats	Features to be selected from the dataset.
size

Returns: Data object with selected features.

◆ setClasses()

template<typename T >

void mltk::Data< T >::setClasses ( const std::vector< int > & classes )

Set the classes to use in the dataset.

Parameters

classes Vector of classes that will be in the dataset.

◆ setClassesAtEnd()

template<typename T = double>

void mltk::Data< T >::setClassesAtEnd ( bool atEnd )

inline

Inform if the classes of the dataset to be loaded are at begining or end of the attributes.

Parameters

atEnd Boolean informing if classes are at end.

◆ setClassesNames()

template<typename T = double>

void mltk::Data< T >::setClassesNames ( const std::vector< std::string > class_names )

inline

Set classes names.

Parameters

class_names strings represeting the classes names.

◆ setDim()

template<typename T >

void mltk::Data< T >::setDim ( size_t dim )

setDim Set the dimension of the points.

Parameters

dim	Dimension to be set.

◆ setFeaturesNames()

template<typename T >

void mltk::Data< T >::setFeaturesNames ( const std::vector< int > & fnames )

setFeaturesNames Set the name of the features of the data.

Parameters

fnames Name of the features.

◆ setIndex()

template<typename T >

void mltk::Data< T >::setIndex ( std::vector< int > index )

Set the index vector for the data.

Parameters

index Index vector.

◆ setName()

template<typename T = double>

void mltk::Data< T >::setName ( const std::string & name )

inline

Set the name of the dataset.

Parameters

name	Name of the dataset.

◆ setPoint()

template<typename T >

void mltk::Data< T >::setPoint	(	int	index,
		std::shared_ptr< Point< T > >	p
	)

setPoint Set the point in a position of the data.

Parameters

index	Index of the point that will be set.
p	Point to be set.

◆ shuffle()

template<typename T >

void mltk::Data< T >::shuffle ( const size_t & seed = 42 )

Shuffle the data with a given seed.

Parameters

seed	Seed given for randomization

◆ size()

template<typename T = double>

size_t mltk::Data< T >::size ( ) const

inline

Returns the size of the dataset.

Returns: int

◆ splitByClasses()

template<typename T >

std::vector< Data< T > > mltk::Data< T >::splitByClasses ( bool keepIndex = false )

Split the dataset by its labels. Only valid for classification datasets.

Returns: Vector containing Data split by classes.

◆ splitSample()

template<typename T >

std::vector< Data< T > > mltk::Data< T >::splitSample	(	const std::size_t &	split_size,
		bool	stratified = `true`,
		bool	keepIndex = `false`,
		size_t	seed = `0`
	)

Split the data by a given size.

Parameters

split_size	Number of samples on each split.
stratified	If true, the split will be done in a stratified manner.
seed	Random generator seed.

Returns: Vector containing the data split.

◆ updatePointValue()

template<typename T >

bool mltk::Data< T >::updatePointValue	(	const size_t &	idx,
		double	value
	)

Updates a Point value.

Parameters

idx	Index of the point to be updated.
value	New value of the point.

Returns: Success of the point value update.

◆ write()

template<typename T >

void mltk::Data< T >::write	(	const std::string &	fname,
		std::string	ext
	)

write Write the data to a file with the given extention.

Parameters

fname	Name of the file.
ext	Extention of the file.

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

Data.hpp

Public Member Functions

Friends

Detailed Description

template<typename T = double> class mltk::Data< T >

Constructor & Destructor Documentation

◆ Data() [1/4]

◆ Data() [2/4]

◆ Data() [3/4]

◆ Data() [4/4]

Member Function Documentation

◆ apply()

◆ changeXVector()

◆ classes()

◆ classesCopy()

◆ classesDistribution()

◆ classesNames()

◆ clear()

◆ computeClassesDistribution()

◆ copy() [1/2]

◆ copy() [2/2]

◆ copyZero()

◆ dim()

◆ getFeature()

◆ getFeaturesNames()

◆ getIndex()

◆ getTime_mult()

◆ getType()

◆ insertFeatures()

◆ insertPoint() [1/3]

◆ insertPoint() [2/3]

◆ insertPoint() [3/3]

◆ isClassification()

◆ isEmpty()

◆ isNormalized()

◆ join()

◆ load() [1/2]

◆ load() [2/2]

◆ name()

◆ normalize()

◆ point()

◆ points() [1/2]

◆ points() [2/2]

◆ removeFeatures() [1/2]

◆ removeFeatures() [2/2]

◆ removePoint()

◆ removePoints()

◆ resetIndex()

◆ sampling()

◆ selectFeatures()

◆ setClasses()

◆ setClassesAtEnd()

◆ setClassesNames()

◆ setDim()

◆ setFeaturesNames()

◆ setIndex()

◆ setName()

◆ setPoint()

◆ shuffle()

◆ size()

◆ splitByClasses()

◆ splitSample()

◆ updatePointValue()

◆ write()

template<typename T = double>
class mltk::Data< T >