SVM (Support Vector Machines) training class for the SVM machine learning. More...

#include <pcl/ml/svm_wrapper.h>

Inheritance diagram for pcl::SVMTrain:

Public Member Functions
	SVMTrain ()
	Constructor.
	~SVMTrain ()
	Destructor.
void	setParameters (SVMParam param)
	Change default training parameters (pcl::SVMParam).
SVMParam	getParameters ()
	Return the current training parameters.
SVMModel	getClassifierModel ()
	Return the result of the training.
void	setInputTrainingSet (std::vector< SVMData > training_set)
	It adds/store the training set with labelled data.
std::vector< SVMData >	getInputTrainingSet ()
	Return the current training set.
void	resetTrainingSet ()
	Reset the training set.
bool	trainClassifier ()
	Start the training of the SVM classifier.
bool	loadProblem (const char *filename)
	Read in a problem (in svmlight format).
void	setDebugMode (bool in)
	Set to 1 for debugging info.
bool	saveTrainingSet (const char *filename)
	Save the raw training set in a file (in svmlight format).
bool	saveNormTrainingSet (const char *filename)
	Save the normalized training set in a file (in svmlight format).
Public Member Functions inherited from pcl::SVM
	SVM ()
	Constructor.
	~SVM ()
	Destructor.
void	getLabel (std::vector< int > &labels)
	Return the labels order from the classifier model.
void	saveClassifierModel (const char *filename)
	Save the classifier model in an extern file (in svmlight format).

Protected Member Functions
void	doCrossValidation ()
	To cross validate the classifier.
void	scaleFactors (std::vector< SVMData > training_set, svm_scaling &scaling)
	It extracts scaling factors from the input training_set.
Protected Member Functions inherited from pcl::SVM
char *	readline (FILE *input)
	To read a line from the input file.
void	exitInputError (int line_num)
	Outputs an error in file reading.
const std::string &	getClassName () const
	Get a string representation of the name of this class.
void	adaptInputToLibSVM (std::vector< SVMData > training_set, svm_problem &prob)
	Convert the input format (vector of SVMData) into a readable format for libSVM.
void	adaptLibSVMToInput (std::vector< SVMData > &training_set, svm_problem prob) const
	Convert the libSVM format (svm_problem) into a easier output format.
bool	loadProblem (const char *filename, svm_problem &prob)
	Load a problem from an extern file.
bool	saveProblem (const char *filename, bool labelled)
	Save the raw problem in an extern file.
bool	saveProblemNorm (const char *filename, svm_problem prob_, bool labelled)
	Save the problem (with normalized values) in an extern file.

Protected Attributes
bool	debug_ {false}
	Set to 1 to see the training output.
int	cross_validation_ {0}
	Set too 1 for cross validating the classifier.
int	nr_fold_ {0}
	Number of folds to be used during cross validation.
std::string	class_name_
bool	labelled_training_set_
char *	line_
int	max_line_len_
SVMModel	model_
SVMParam	param_
svm_problem	prob_
svm_scaling	scaling_
std::vector< SVMData >	training_set_
Protected Attributes inherited from pcl::SVM
std::vector< SVMData >	training_set_
svm_problem	prob_
SVMModel	model_
svm_scaling	scaling_
SVMParam	param_
std::string	class_name_
char *	line_ {nullptr}
int	max_line_len_ {10000}
bool	labelled_training_set_

Additional Inherited Members
Static Protected Member Functions inherited from pcl::SVM
static void	printNull (const char *)
	Set for output printings during classification.

Detailed Description

SVM (Support Vector Machines) training class for the SVM machine learning.

It creates a model for the classifier from a labelled input dataset.

OPTIONAL: pcl::SVMParam has to be given as input to vary the default training method and parameters.

Definition at line 235 of file svm_wrapper.h.

Constructor & Destructor Documentation

◆ SVMTrain()

pcl::SVMTrain::SVMTrain ( )

inline

Constructor.

Definition at line 268 of file svm_wrapper.h.

References class_name_, and pcl::SVM::printNull().

◆ ~SVMTrain()

pcl::SVMTrain::~SVMTrain ( )

inline

Destructor.

Definition at line 276 of file svm_wrapper.h.

References model_.

Member Function Documentation

◆ doCrossValidation()

void pcl::SVMTrain::doCrossValidation ( )

protected

To cross validate the classifier.

It is automatic for probability estimate.

◆ getClassifierModel()

SVMModel pcl::SVMTrain::getClassifierModel ( )

inline

Return the result of the training.

Definition at line 298 of file svm_wrapper.h.

References model_.

◆ getInputTrainingSet()

std::vector< SVMData > pcl::SVMTrain::getInputTrainingSet ( )

inline

Return the current training set.

Definition at line 312 of file svm_wrapper.h.

References training_set_.

◆ getParameters()

SVMParam pcl::SVMTrain::getParameters ( )

inline

Return the current training parameters.

Definition at line 291 of file svm_wrapper.h.

References param_.

◆ loadProblem()

bool pcl::SVMTrain::loadProblem ( const char * filename )

inline

Read in a problem (in svmlight format).

Returns: false if fails

Definition at line 336 of file svm_wrapper.h.

References pcl::SVM::loadProblem(), and prob_.

◆ resetTrainingSet()

void pcl::SVMTrain::resetTrainingSet ( )

inline

Reset the training set.

Definition at line 319 of file svm_wrapper.h.

References training_set_.

◆ saveNormTrainingSet()

bool pcl::SVMTrain::saveNormTrainingSet ( const char * filename )

inline

Save the normalized training set in a file (in svmlight format).

Returns: false if fails

Definition at line 368 of file svm_wrapper.h.

References prob_, and pcl::SVM::saveProblemNorm().

◆ saveTrainingSet()

bool pcl::SVMTrain::saveTrainingSet ( const char * filename )

inline

Save the raw training set in a file (in svmlight format).

Returns: false if fails

Definition at line 358 of file svm_wrapper.h.

References pcl::SVM::saveProblem().

◆ scaleFactors()

void pcl::SVMTrain::scaleFactors	(	std::vector< SVMData >	training_set,
		svm_scaling &	scaling )

protected

It extracts scaling factors from the input training_set.

The scaling of the training_set is a mandatory for a good training of the classifier.

◆ setDebugMode()

void pcl::SVMTrain::setDebugMode ( bool in )

inline

Set to 1 for debugging info.

Definition at line 343 of file svm_wrapper.h.

References debug_, and pcl::SVM::printNull().

◆ setInputTrainingSet()

void pcl::SVMTrain::setInputTrainingSet ( std::vector< SVMData > training_set )

inline

It adds/store the training set with labelled data.

Definition at line 305 of file svm_wrapper.h.

References training_set_.

◆ setParameters()

void pcl::SVMTrain::setParameters ( SVMParam param )

inline

Change default training parameters (pcl::SVMParam).

Definition at line 284 of file svm_wrapper.h.

References param_.

◆ trainClassifier()

bool pcl::SVMTrain::trainClassifier ( )

Start the training of the SVM classifier.

Returns: false if fails

Member Data Documentation

◆ class_name_

std::string pcl::SVM::class_name_

protected

Definition at line 130 of file svm_wrapper.h.

Referenced by SVMTrain().

◆ cross_validation_

int pcl::SVMTrain::cross_validation_ {0}

protected

Set too 1 for cross validating the classifier.

Definition at line 250 of file svm_wrapper.h.

◆ debug_

bool pcl::SVMTrain::debug_ {false}

protected

Set to 1 to see the training output.

Definition at line 248 of file svm_wrapper.h.

Referenced by setDebugMode().

◆ labelled_training_set_

bool pcl::SVM::labelled_training_set_

protected

Definition at line 134 of file svm_wrapper.h.

◆ line_

char* pcl::SVM::line_

protected

Definition at line 132 of file svm_wrapper.h.

◆ max_line_len_

int pcl::SVM::max_line_len_

protected

Definition at line 133 of file svm_wrapper.h.

◆ model_

SVMModel pcl::SVM::model_

protected

Definition at line 126 of file svm_wrapper.h.

Referenced by getClassifierModel(), and ~SVMTrain().

◆ nr_fold_

int pcl::SVMTrain::nr_fold_ {0}

protected

Number of folds to be used during cross validation.

It indicates in how many parts is split the input training set.

Definition at line 253 of file svm_wrapper.h.

◆ param_

SVMParam pcl::SVM::param_

protected

Definition at line 129 of file svm_wrapper.h.

Referenced by getParameters(), and setParameters().

◆ prob_

svm_problem pcl::SVM::prob_

protected

Definition at line 125 of file svm_wrapper.h.

Referenced by loadProblem(), and saveNormTrainingSet().

◆ scaling_

svm_scaling pcl::SVM::scaling_

protected

Definition at line 127 of file svm_wrapper.h.

◆ training_set_

std::vector<SVMData> pcl::SVM::training_set_

protected

Definition at line 124 of file svm_wrapper.h.

Referenced by getInputTrainingSet(), resetTrainingSet(), and setInputTrainingSet().

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

pcl/ml/svm_wrapper.h

Public Member Functions

Protected Member Functions

Protected Attributes

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

◆ SVMTrain()

◆ ~SVMTrain()

Member Function Documentation

◆ doCrossValidation()

◆ getClassifierModel()

◆ getInputTrainingSet()

◆ getParameters()

◆ loadProblem()

◆ resetTrainingSet()

◆ saveNormTrainingSet()

◆ saveTrainingSet()

◆ scaleFactors()

◆ setDebugMode()

◆ setInputTrainingSet()

◆ setParameters()

◆ trainClassifier()

Member Data Documentation

◆ class_name_

◆ cross_validation_

◆ debug_

◆ labelled_training_set_

◆ line_

◆ max_line_len_

◆ model_

◆ nr_fold_

◆ param_

◆ prob_

◆ scaling_

◆ training_set_