OpenCLTemplate.MachineLearning.MultiClassSVM Class Reference

Multiple training SVM. More...

Public Member Functions
	MultiClassSVM (TrainingSet TSet)
	Creates a new multiclass SVM using desired outputs from training set. Classifications -1.0f are negative for all sets.
	MultiClassSVM (TrainingSet TSet, ProblemConfig SVMCfg)
	Creates a new multiclass SVM using desired outputs from training set. Classifications -1.0f are negative for all sets.
void	Train ()
	Trains all SVMs in this multiclass SVM.
void	Train (float tolPositive, float tolNegative)
	Trains all SVMs in this multiclass SVM precalibrating kernels.
float	TrainWithCrossValidation (TrainingSet CrossValidationSet)
	Trains current SVM with cross-validation, adjusting kernel parameter lambda and box parameter C. Returns best achieved efficiency.
float	TrainWithCrossValidation (TrainingSet CrossValidationSet, float[] LambdaSet, float[] CSet)
	Trains current SVM with cross-validation, adjusting kernel parameter lambda and box parameter C. Returns best achieved efficiency.
float	ClassifyWithRejection (TrainingUnit Sample)
	Attempts to classify a sample within a given category. Returns -1 if no classification was achieved.
float	Classify (TrainingUnit Sample, out float maxVal)
	Classifies a sample within a given category even if all SVMs predict it doesn`t belong to any.
float[]	Classify (CLCalc.Program.Image2D Samples, out float[] maxVals)
	Classifies a given set of Samples (image2d of floats) each one in a category. Each row of the image is a sample to be classified and the features should be stored in the columns. The number of columns Ncol = Nfeatures/4 since each pixel holds 4 floats.
float	GetHitRate (TrainingSet TestSet)
	Gets SVM hit rate.
float	GetInternalHitRate ()
	Gets average internal hit rate.
	MultiClassSVM (TrainingSet TSet)
	Creates a new multiclass SVM using desired outputs from training set. Classifications -1.0f are negative for all sets.
	MultiClassSVM (TrainingSet TSet, ProblemConfig SVMCfg, bool PreCalibrate)
	Creates a new multiclass SVM using desired outputs from training set. Classifications -1.0f are negative for all sets.
void	Train ()
	Trains all SVMs in this multiclass SVM.
void	Train (float tolPositive, float tolNegative)
	Trains all SVMs in this multiclass SVM precalibrating kernels.
float	TrainWithCrossValidation (TrainingSet CrossValidationSet)
	Trains current SVM with cross-validation, adjusting kernel parameter lambda and box parameter C. Returns best achieved efficiency.
float	TrainWithCrossValidation (TrainingSet CrossValidationSet, float[] LambdaSet, float[] CSet)
	Trains current SVM with cross-validation, adjusting kernel parameter lambda and box parameter C. Returns best achieved efficiency.
float	ClassifyWithRejection (TrainingUnit Sample)
	Attempts to classify a sample within a given category. Returns -1 if no classification was achieved.
float	Classify (TrainingUnit Sample, out float maxVal)
	Classifies a sample within a given category even if all SVMs predict it doesn`t belong to any.
float[]	Classify (CLCalc.Program.Image2D Samples, out float[] maxVals)
	Classifies a given set of Samples (image2d of floats) each one in a category. Each row of the image is a sample to be classified and the features should be stored in the columns. The number of columns Ncol = Nfeatures/4 since each pixel holds 4 floats.
float	GetHitRate (TrainingSet TestSet)
	Gets SVM hit rate.
float	GetInternalHitRate ()
	Gets average internal hit rate.
Static Public Member Functions
static TrainingSet	GetCrossValidationSet (TrainingSet Set, float CrossValidationSetPercent)
	Extracts a cross validation set from a given set.
static TrainingSet	GetCrossValidationSet (TrainingSet Set, float CrossValidationSetPercent)
	Extracts a cross validation set from a given set.
Public Attributes
List< float >	Classifications
	List of possible classifications.
List< SVM >	SVMs
	SVMs to perform each classification.

Detailed Description

Multiple training SVM.

Constructor & Destructor Documentation

OpenCLTemplate.MachineLearning.MultiClassSVM.MultiClassSVM ( TrainingSet TSet )

Creates a new multiclass SVM using desired outputs from training set. Classifications -1.0f are negative for all sets.

Parameters:

TSet

Training set

OpenCLTemplate.MachineLearning.MultiClassSVM.MultiClassSVM	(	TrainingSet	TSet,
		ProblemConfig	SVMCfg
	)

Creates a new multiclass SVM using desired outputs from training set. Classifications -1.0f are negative for all sets.

Parameters:

	TSet	Training set
	SVMCfg	Configuration parameters

OpenCLTemplate.MachineLearning.MultiClassSVM.MultiClassSVM ( TrainingSet TSet )

Creates a new multiclass SVM using desired outputs from training set. Classifications -1.0f are negative for all sets.

Parameters:

TSet

Training set

OpenCLTemplate.MachineLearning.MultiClassSVM.MultiClassSVM	(	TrainingSet	TSet,
		ProblemConfig	SVMCfg,
		bool	PreCalibrate
	)

Creates a new multiclass SVM using desired outputs from training set. Classifications -1.0f are negative for all sets.

Parameters:

	TSet	Training set
	SVMCfg	Configuration parameters
	PreCalibrate	Precalibrate RBF parameter lambda? This will ignore the given value

Member Function Documentation

float [] OpenCLTemplate.MachineLearning.MultiClassSVM.Classify	(	CLCalc.Program.Image2D	Samples,
		out float[]	maxVals
	)

Classifies a given set of Samples (image2d of floats) each one in a category. Each row of the image is a sample to be classified and the features should be stored in the columns. The number of columns Ncol = Nfeatures/4 since each pixel holds 4 floats.

Parameters:

	Samples	Image2D containing samples to be classified
	maxVals	Maximum values found

Returns:

float OpenCLTemplate.MachineLearning.MultiClassSVM.Classify	(	TrainingUnit	Sample,
		out float	maxVal
	)

Classifies a sample within a given category even if all SVMs predict it doesn`t belong to any.

Parameters:

	Sample	Sample to classify
	maxVal	Maximum classification value found

float [] OpenCLTemplate.MachineLearning.MultiClassSVM.Classify	(	CLCalc.Program.Image2D	Samples,
		out float[]	maxVals
	)

Classifies a given set of Samples (image2d of floats) each one in a category. Each row of the image is a sample to be classified and the features should be stored in the columns. The number of columns Ncol = Nfeatures/4 since each pixel holds 4 floats.

Parameters:

	Samples	Image2D containing samples to be classified
	maxVals	Maximum values found

Returns:

float OpenCLTemplate.MachineLearning.MultiClassSVM.Classify	(	TrainingUnit	Sample,
		out float	maxVal
	)

Classifies a sample within a given category even if all SVMs predict it doesn`t belong to any.

Parameters:

	Sample	Sample to classify
	maxVal	Maximum classification value found

float OpenCLTemplate.MachineLearning.MultiClassSVM.ClassifyWithRejection ( TrainingUnit Sample )

Attempts to classify a sample within a given category. Returns -1 if no classification was achieved.

float OpenCLTemplate.MachineLearning.MultiClassSVM.ClassifyWithRejection ( TrainingUnit Sample )

Attempts to classify a sample within a given category. Returns -1 if no classification was achieved.

static TrainingSet OpenCLTemplate.MachineLearning.MultiClassSVM.GetCrossValidationSet	(	TrainingSet	Set,
		float	CrossValidationSetPercent
	)			`[static]`

Extracts a cross validation set from a given set.

Parameters:

	Set	Set to extract cross validation from
	CrossValidationSetPercent	Percentage of elements to extract

static TrainingSet OpenCLTemplate.MachineLearning.MultiClassSVM.GetCrossValidationSet	(	TrainingSet	Set,
		float	CrossValidationSetPercent
	)			`[static]`

Extracts a cross validation set from a given set.

Parameters:

	Set	Set to extract cross validation from
	CrossValidationSetPercent	Percentage of elements to extract

float OpenCLTemplate.MachineLearning.MultiClassSVM.GetHitRate ( TrainingSet TestSet )

Gets SVM hit rate.

Parameters:

TestSet

Test set

float OpenCLTemplate.MachineLearning.MultiClassSVM.GetHitRate ( TrainingSet TestSet )

Gets SVM hit rate.

Parameters:

TestSet

Test set

float OpenCLTemplate.MachineLearning.MultiClassSVM.GetInternalHitRate ( )

Gets average internal hit rate.

float OpenCLTemplate.MachineLearning.MultiClassSVM.GetInternalHitRate ( )

Gets average internal hit rate.

void OpenCLTemplate.MachineLearning.MultiClassSVM.Train	(	float	tolPositive,
		float	tolNegative
	)

Trains all SVMs in this multiclass SVM precalibrating kernels.

Parameters:

	tolPositive	Positive kernels average should be greater than tolPositive
	tolNegative	Negative kernels average should be lesser than tolNegative

void OpenCLTemplate.MachineLearning.MultiClassSVM.Train ( )

Trains all SVMs in this multiclass SVM.

void OpenCLTemplate.MachineLearning.MultiClassSVM.Train	(	float	tolPositive,
		float	tolNegative
	)

Trains all SVMs in this multiclass SVM precalibrating kernels.

Parameters:

	tolPositive	Positive kernels average should be greater than tolPositive
	tolNegative	Negative kernels average should be lesser than tolNegative

void OpenCLTemplate.MachineLearning.MultiClassSVM.Train ( )

Trains all SVMs in this multiclass SVM.

float OpenCLTemplate.MachineLearning.MultiClassSVM.TrainWithCrossValidation	(	TrainingSet	CrossValidationSet,
		float[]	LambdaSet,
		float[]	CSet
	)

Trains current SVM with cross-validation, adjusting kernel parameter lambda and box parameter C. Returns best achieved efficiency.

Parameters:

	CrossValidationSet	Cross validation set
	LambdaSet	Lambda set
	CSet	C values set

float OpenCLTemplate.MachineLearning.MultiClassSVM.TrainWithCrossValidation ( TrainingSet CrossValidationSet )

Trains current SVM with cross-validation, adjusting kernel parameter lambda and box parameter C. Returns best achieved efficiency.

Parameters:

CrossValidationSet

Cross validation set

float OpenCLTemplate.MachineLearning.MultiClassSVM.TrainWithCrossValidation	(	TrainingSet	CrossValidationSet,
		float[]	LambdaSet,
		float[]	CSet
	)

Trains current SVM with cross-validation, adjusting kernel parameter lambda and box parameter C. Returns best achieved efficiency.

Parameters:

	CrossValidationSet	Cross validation set
	LambdaSet	Lambda set
	CSet	C values set

float OpenCLTemplate.MachineLearning.MultiClassSVM.TrainWithCrossValidation ( TrainingSet CrossValidationSet )

Trains current SVM with cross-validation, adjusting kernel parameter lambda and box parameter C. Returns best achieved efficiency.

Parameters:

CrossValidationSet

Cross validation set

Member Data Documentation

List< float > OpenCLTemplate::MachineLearning.MultiClassSVM::Classifications

List of possible classifications.

List< SVM > OpenCLTemplate::MachineLearning.MultiClassSVM::SVMs

SVMs to perform each classification.

OpenCLTemplate.MachineLearning.MultiClassSVM Class Reference

Public Member Functions

Static Public Member Functions

Public Attributes

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation

Member Data Documentation