Inheritance diagram for SVM:

Public Types
enum	SVMTypes { C_SVC = 0, NU_SVC, ONE_CLASS, EPSILON_SVR, NU_SVR }

enum	SVMKernelTypes { LINEAR_KERNEL = 0, POLY_KERNEL, RBF_KERNEL, SIGMOID_KERNEL, PRECOMPUTED_KERNEL }

Public Types inherited from Classifier
typedef std::map< std::string, Classifier ()() >	StringClassifierMap

Public Types inherited from MLBase
enum	BaseTypes { BASE_TYPE_NOT_SET =0, CLASSIFIER, REGRESSIFIER, CLUSTERER }

Public Member Functions
	SVM (UINT kernelType=LINEAR_KERNEL, UINT svmType=C_SVC, bool useScaling=true, bool useNullRejection=false, bool useAutoGamma=true, Float gamma=0.1, UINT degree=3, Float coef0=0, Float nu=0.5, Float C=1, bool useCrossValidation=false, UINT kFoldValue=10)

	SVM (const SVM &rhs)

virtual	~SVM ()

SVM &	operator= (const SVM &rhs)

virtual bool	deepCopyFrom (const Classifier *classifier)

virtual bool	train_ (ClassificationData &trainingData)

virtual bool	predict_ (VectorFloat &inputVector)

virtual bool	clear ()

virtual bool	saveModelToFile (std::fstream &file) const

virtual bool	loadModelFromFile (std::fstream &file)

bool	init (UINT kernelType, UINT svmType, bool useScaling, bool useNullRejection, bool useAutoGamma, Float gamma, UINT degree, Float coef0, Float nu, Float C, bool useCrossValidation, UINT kFoldValue)

void	initDefaultSVMSettings ()

bool	getIsCrossValidationTrainingEnabled () const

bool	getIsAutoGammaEnabled () const

std::string	getSVMType () const

std::string	getKernelType () const

UINT	getDegree () const

virtual UINT	getNumClasses () const

Float	getGamma () const

Float	getNu () const

Float	getCoef0 () const

Float	getC () const

Float	getCrossValidationResult () const

struct svm_model *	getModel () const

bool	setSVMType (const UINT svmType)

bool	setKernelType (const UINT kernelType)

bool	setGamma (const Float gamma)

bool	setDegree (const UINT degree)

bool	setNu (const Float nu)

bool	setCoef0 (const Float coef0)

bool	setC (const Float C)

bool	setKFoldCrossValidationValue (const UINT kFoldValue)

bool	enableAutoGamma (const bool useAutoGamma)

bool	enableCrossValidationTraining (const bool useCrossValidation)

Public Member Functions inherited from Classifier
	Classifier (void)

virtual	~Classifier (void)

bool	copyBaseVariables (const Classifier *classifier)

virtual bool	reset ()

std::string	getClassifierType () const

bool	getSupportsNullRejection () const

bool	getNullRejectionEnabled () const

Float	getNullRejectionCoeff () const

Float	getMaximumLikelihood () const

Float	getBestDistance () const

Float	getPhase () const

UINT	getClassLabelIndexValue (UINT classLabel) const

UINT	getPredictedClassLabel () const

VectorFloat	getClassLikelihoods () const

VectorFloat	getClassDistances () const

VectorFloat	getNullRejectionThresholds () const

Vector< UINT >	getClassLabels () const

Vector< MinMax >	getRanges () const

bool	enableNullRejection (bool useNullRejection)

virtual bool	setNullRejectionCoeff (Float nullRejectionCoeff)

virtual bool	setNullRejectionThresholds (VectorFloat newRejectionThresholds)

virtual bool	recomputeNullRejectionThresholds ()

bool	getTimeseriesCompatible () const

Classifier *	createNewInstance () const

Classifier *	deepCopy () const

const Classifier *	getClassifierPointer () const

const Classifier &	getBaseClassifier () const

Public Member Functions inherited from MLBase
	MLBase (void)

virtual	~MLBase (void)

bool	copyMLBaseVariables (const MLBase *mlBase)

virtual bool	train (ClassificationData trainingData)

virtual bool	train (RegressionData trainingData)

virtual bool	train_ (RegressionData &trainingData)

virtual bool	train (TimeSeriesClassificationData trainingData)

virtual bool	train_ (TimeSeriesClassificationData &trainingData)

virtual bool	train (ClassificationDataStream trainingData)

virtual bool	train_ (ClassificationDataStream &trainingData)

virtual bool	train (UnlabelledData trainingData)

virtual bool	train_ (UnlabelledData &trainingData)

virtual bool	train (MatrixFloat data)

virtual bool	train_ (MatrixFloat &data)

virtual bool	predict (VectorFloat inputVector)

virtual bool	predict (MatrixFloat inputMatrix)

virtual bool	predict_ (MatrixFloat &inputMatrix)

virtual bool	map (VectorFloat inputVector)

virtual bool	map_ (VectorFloat &inputVector)

virtual bool	print () const

virtual bool	save (const std::string filename) const

virtual bool	load (const std::string filename)

virtual bool	saveModelToFile (std::string filename) const

virtual bool	loadModelFromFile (std::string filename)

virtual bool	getModel (std::ostream &stream) const

Float	scale (const Float &x, const Float &minSource, const Float &maxSource, const Float &minTarget, const Float &maxTarget, const bool constrain=false)

virtual std::string	getModelAsString () const

DataType	getInputType () const

DataType	getOutputType () const

UINT	getBaseType () const

UINT	getNumInputFeatures () const

UINT	getNumInputDimensions () const

UINT	getNumOutputDimensions () const

UINT	getMinNumEpochs () const

UINT	getMaxNumEpochs () const

UINT	getValidationSetSize () const

UINT	getNumTrainingIterationsToConverge () const

Float	getMinChange () const

Float	getLearningRate () const

Float	getRootMeanSquaredTrainingError () const

Float	getTotalSquaredTrainingError () const

Float	getValidationSetAccuracy () const

VectorFloat	getValidationSetPrecision () const

VectorFloat	getValidationSetRecall () const

bool	getUseValidationSet () const

bool	getRandomiseTrainingOrder () const

bool	getTrained () const

bool	getModelTrained () const

bool	getScalingEnabled () const

bool	getIsBaseTypeClassifier () const

bool	getIsBaseTypeRegressifier () const

bool	getIsBaseTypeClusterer () const

bool	enableScaling (const bool useScaling)

bool	setMaxNumEpochs (const UINT maxNumEpochs)

bool	setMinNumEpochs (const UINT minNumEpochs)

bool	setMinChange (const Float minChange)

bool	setLearningRate (const Float learningRate)

bool	setUseValidationSet (const bool useValidationSet)

bool	setValidationSetSize (const UINT validationSetSize)

bool	setRandomiseTrainingOrder (const bool randomiseTrainingOrder)

bool	setTrainingLoggingEnabled (const bool loggingEnabled)

bool	registerTrainingResultsObserver (Observer< TrainingResult > &observer)

bool	registerTestResultsObserver (Observer< TestInstanceResult > &observer)

bool	removeTrainingResultsObserver (const Observer< TrainingResult > &observer)

bool	removeTestResultsObserver (const Observer< TestInstanceResult > &observer)

bool	removeAllTrainingObservers ()

bool	removeAllTestObservers ()

bool	notifyTrainingResultsObservers (const TrainingResult &data)

bool	notifyTestResultsObservers (const TestInstanceResult &data)

MLBase *	getMLBasePointer ()

const MLBase *	getMLBasePointer () const

Vector< TrainingResult >	getTrainingResults () const

Public Member Functions inherited from GRTBase
	GRTBase (void)

virtual	~GRTBase (void)

bool	copyGRTBaseVariables (const GRTBase *GRTBase)

std::string	getClassType () const

std::string	getLastWarningMessage () const

std::string	getLastErrorMessage () const

std::string	getLastInfoMessage () const

bool	setInfoLoggingEnabled (const bool loggingEnabled)

bool	setWarningLoggingEnabled (const bool loggingEnabled)

bool	setErrorLoggingEnabled (const bool loggingEnabled)

GRTBase *	getGRTBasePointer ()

const GRTBase *	getGRTBasePointer () const

Public Member Functions inherited from Observer< TrainingResult >
virtual void	notify (const TrainingResult &data)

Public Member Functions inherited from Observer< TestInstanceResult >
virtual void	notify (const TestInstanceResult &data)

Protected Member Functions
void	deleteProblemSet ()

bool	validateProblemAndParameters ()

bool	validateSVMType (UINT svmType)

bool	validateKernelType (UINT kernelType)

bool	convertClassificationDataToLIBSVMFormat (ClassificationData &trainingData)

bool	trainSVM ()

bool	predictSVM (VectorFloat &inputVector)

bool	predictSVM (VectorFloat &inputVector, Float &maxProbability, VectorFloat &probabilites)

bool	loadLegacyModelFromFile (std::fstream &file)

struct svm_model *	deepCopyModel () const

bool	deepCopyProblem (const struct svm_problem &source_problem, struct svm_problem &target_problem, const unsigned int numInputDimensions) const

bool	deepCopyParam (const svm_parameter &source_param, svm_parameter &target_param) const

Protected Member Functions inherited from Classifier
bool	saveBaseSettingsToFile (std::fstream &file) const

bool	loadBaseSettingsFromFile (std::fstream &file)

Protected Member Functions inherited from MLBase
bool	saveBaseSettingsToFile (std::fstream &file) const

bool	loadBaseSettingsFromFile (std::fstream &file)

Protected Member Functions inherited from GRTBase
Float	SQR (const Float &x) const

Protected Attributes
bool	problemSet

struct svm_model *	model

struct svm_parameter	param

struct svm_problem	prob

UINT	kFoldValue

Float	classificationThreshold

Float	crossValidationResult

bool	useAutoGamma

bool	useCrossValidation

Protected Attributes inherited from Classifier
std::string	classifierType

bool	supportsNullRejection

bool	useNullRejection

UINT	numClasses

UINT	predictedClassLabel

UINT	classifierMode

Float	nullRejectionCoeff

Float	maxLikelihood

Float	bestDistance

Float	phase

VectorFloat	classLikelihoods

VectorFloat	classDistances

VectorFloat	nullRejectionThresholds

Vector< UINT >	classLabels

Vector< MinMax >	ranges

Protected Attributes inherited from MLBase
bool	trained

bool	useScaling

DataType	inputType

DataType	outputType

UINT	baseType

UINT	numInputDimensions

UINT	numOutputDimensions

UINT	numTrainingIterationsToConverge

UINT	minNumEpochs

UINT	maxNumEpochs

UINT	validationSetSize

Float	learningRate

Float	minChange

Float	rootMeanSquaredTrainingError

Float	totalSquaredTrainingError

Float	validationSetAccuracy

bool	useValidationSet

bool	randomiseTrainingOrder

VectorFloat	validationSetPrecision

VectorFloat	validationSetRecall

Random	random

std::vector< TrainingResult >	trainingResults

TrainingResultsObserverManager	trainingResultsObserverManager

TestResultsObserverManager	testResultsObserverManager

Protected Attributes inherited from GRTBase
std::string	classType

DebugLog	debugLog

ErrorLog	errorLog

InfoLog	infoLog

TrainingLog	trainingLog

TestingLog	testingLog

WarningLog	warningLog

Static Protected Attributes
static RegisterClassifierModule< SVM >	registerModule

Additional Inherited Members
Static Public Member Functions inherited from Classifier
static Classifier *	createInstanceFromString (std::string const &classifierType)

static Vector< std::string >	getRegisteredClassifiers ()

Static Public Member Functions inherited from GRTBase
static std::string	getGRTVersion (bool returnRevision=true)

static std::string	getGRTRevison ()

Protected Types inherited from Classifier
enum	ClassifierModes { STANDARD_CLASSIFIER_MODE =0, TIMESERIES_CLASSIFIER_MODE }

Static Protected Member Functions inherited from Classifier
static StringClassifierMap *	getMap ()

Detailed Description

Definition at line 49 of file SVM.h.

Constructor & Destructor Documentation

SVM::SVM	(	UINT	kernelType = `LINEAR_KERNEL`,
		UINT	svmType = `C_SVC`,
		bool	useScaling = `true`,
		bool	useNullRejection = `false`,
		bool	useAutoGamma = `true`,
		Float	gamma = `0.1`,
		UINT	degree = `3`,
		Float	coef0 = `0`,
		Float	nu = `0.5`,
		Float	C = `1`,
		bool	useCrossValidation = `false`,
		UINT	kFoldValue = `10`
	)

Default constructor.

Set the initial SVM settings, although these can be changed at any time using either init(...) function of the set... functions.

Parameters

kernelType	this sets the SVM kernelType. Options are LINEAR_KERNEL, POLY_KERNEL, RBF_KERNEL, SIGMOID_KERNEL, PRECOMPUTED_KERNEL. The default kernelType is kernelType=LINEAR_KERNEL
svmType	this sets the SVM type. Options are C_SVC, NU_SVC, ONE_CLASS, EPSILON_SVR, NU_SVR. The default svmType is svmType=C_SVC
useScaling	sets if the training/prediction data will be scaled to the default range of [-1. 1.]. The SVM algorithm commonly achieves a better classification result if scaling is turned on. The default useScaling value is useScaling=true
useNullRejection	sets if a predicted class will be rejected if the classes' probability is below the classificationThreshold. The default value is useNullRejection=false
useAutoGamma	sets if the SVM gamma parameter will automatically be computed, if set to true then gamma will be set to (1.0/numFeatures), where numFeatures is the number of features in the training data. The default value is useAutoGamma=true
gamma	sets the SVM gamma parameter. The default value is gamma=0.1
degree	sets the SVM degree parameter. The default value is degree=3
coef0	sets the SVM coef0 parameter. The default value is coef0=0
nu	sets the SVM nu parameter. The default value is nu=0.5
C	sets the SVM C parameter. The default value is C=1
useCrossValidation	sets if the SVM model will be trained using cross validation. The default value is useCrossValidation=false
kFoldValue	sets the number of folds that will be used for cross validation. The default value is kFoldValue=10

Definition at line 28 of file SVM.cpp.

SVM::SVM ( const SVM & rhs )

Default copy constructor. Copies the settings from the rhs SVM instance to this instance

Parameters

rhs	another instance of a SVM

Definition at line 72 of file SVM.cpp.

SVM::~SVM ( )

virtual

Default Destructor

Definition at line 90 of file SVM.cpp.

Member Function Documentation

bool SVM::clear ( )

virtual

Clears any previous model or problem.

Reimplemented from Classifier.

Definition at line 932 of file SVM.cpp.

bool SVM::deepCopyFrom ( const Classifier * classifier )

virtual

This is required for the Gesture Recognition Pipeline for when the pipeline.setClassifier(...) method is called. It clones the data from the Base Class Classifier pointer (which should be pointing to an SVM instance) into this instance

Parameters

classifier a pointer to the Classifier Base Class, this should be pointing to another SVM instance

Returns: returns true if the clone was successfull, false otherwise

Reimplemented from Classifier.

Definition at line 116 of file SVM.cpp.

bool SVM::enableAutoGamma ( const bool useAutoGamma )

Sets if the gamma parameter will be automatically computed from the training data.

Parameters

useAutoGamma the new useAutoGamma setting return returns true if the useAutoGamma was set, false otherwise

Definition at line 1111 of file SVM.cpp.

bool SVM::enableCrossValidationTraining ( const bool useCrossValidation )

Sets if cross validation should be used during the training phase.

Parameters

useCrossValidation the new useCrossValidation setting return returns true if the useCrossValidation was set, false otherwise

Definition at line 1116 of file SVM.cpp.

Float SVM::getC ( ) const

Gets the current C value.

Returns: returns the current C value.

Definition at line 1047 of file SVM.cpp.

Float SVM::getCoef0 ( ) const

Gets the current coef0 value.

Returns: returns the current coef0 value.

Definition at line 1040 of file SVM.cpp.

Float SVM::getCrossValidationResult ( ) const

Gets the last cross validation result, if the model has been trained and cross validation was enabled.

Returns: returns the last cross validation result.

Definition at line 1054 of file SVM.cpp.

UINT SVM::getDegree ( ) const

Gets the current degree value.

Returns: returns the current degree value.

Definition at line 1019 of file SVM.cpp.

Float SVM::getGamma ( ) const

Gets the current gamma value.

Returns: returns the current gamma value.

Definition at line 1026 of file SVM.cpp.

bool SVM::getIsAutoGammaEnabled ( ) const

Gets if the gamma parameter will be automatically computed from the training data. If false, then the user needs to set the gamma parameter.

Returns: returns true if the gamma parameter will be automatically computed, false otherwise

Definition at line 950 of file SVM.cpp.

bool SVM::getIsCrossValidationTrainingEnabled ( ) const

Gets if cross validation is enabled.

Returns: returns true if cross validation is enabled, false otherwise

Definition at line 946 of file SVM.cpp.

std::string SVM::getKernelType ( ) const

Gets the current kernel type as a string.

If the type has not been set correctly then the string value UNKNOWN will be returned.

Returns: returns the current kernel type as a string.

Definition at line 985 of file SVM.cpp.

Float SVM::getNu ( ) const

Gets the current nu value.

Returns: returns the current nu value.

Definition at line 1033 of file SVM.cpp.

UINT SVM::getNumClasses ( ) const

virtual

Returns the number of classes in the trained model.

If the model has not been trained then 0 will be returned.

Returns: returns the number of classes in the trained model or 0 if the model has not been trained

Reimplemented from Classifier.

Definition at line 1014 of file SVM.cpp.

std::string SVM::getSVMType ( ) const

Gets the current SVM type as a string.

If the type has not been set correctly then the string value UNKNOWN will be returned.

Returns: returns the current SVM type as a string.

Definition at line 954 of file SVM.cpp.

bool SVM::init	(	UINT	kernelType,
		UINT	svmType,
		bool	useScaling,
		bool	useNullRejection,
		bool	useAutoGamma,
		Float	gamma,
		UINT	degree,
		Float	coef0,
		Float	nu,
		Float	C,
		bool	useCrossValidation,
		UINT	kFoldValue
	)

This initializes the SVM settings and parameters. Any previous model, settings, or problems will be cleared.

Parameters

kernelType	this sets the SVM kernelType. Options are LINEAR_KERNEL, POLY_KERNEL, RBF_KERNEL, SIGMOID_KERNEL, PRECOMPUTED_KERNEL
UINT	svmType: this sets the SVM type. Options are C_SVC, NU_SVC, ONE_CLASS, EPSILON_SVR, NU_SVR
useScaling	sets if the training/prediction data will be scaled to the default range of [-1. 1.]. The SVM algorithm commonly achieves a better classification result if scaling is turned on
useNullRejection	sets if a predicted class will be rejected if the classes' probability is below the classificationThreshold
useAutoGamma	sets if the SVM gamma parameter will automatically be computed, if set to true then gamma will be set to (1.0/numFeatures), where numFeatures is the number of features in the training data
gamma	sets the SVM gamma parameter
degree	sets the SVM degree parameter
coef0	sets the SVM coef0 parameter
nu	sets the SVM nu parameter
C	sets the SVM C parameter
useCrossValidation	sets if the SVM model will be trained using cross validation
kFoldValue	sets the number of folds that will be used for cross validation

Returns: returns true if the SVM was initialized, false otherwise

Definition at line 199 of file SVM.cpp.

void SVM::initDefaultSVMSettings ( )

This initializes the default SVM settings and parameters. Any previous model, settings, or problems will be cleared.

Definition at line 255 of file SVM.cpp.

bool SVM::loadLegacyModelFromFile ( std::fstream & file )

protected

Load the ranges

Definition at line 1338 of file SVM.cpp.

bool SVM::loadModelFromFile ( std::fstream & file )

virtual

This loads a trained SVM model from a file. This overrides the loadModelFromFile function in the Classifier base class.

Parameters

file	a reference to the file the SVM model will be loaded from

Returns: returns true if the model was loaded successfully, false otherwise

Reimplemented from MLBase.

Definition at line 627 of file SVM.cpp.

SVM & SVM::operator= ( const SVM & rhs )

Defines how the data from the rhs SVM should be copied to this SVM

Parameters

rhs	another instance of a SVM

Returns: returns a pointer to this instance of the SVM

Definition at line 94 of file SVM.cpp.

bool SVM::predict_ ( VectorFloat & inputVector )

virtual

This predicts the class of the inputVector. This overrides the predict function in the Classifier base class.

Parameters

inputVector the input vector to classify

Returns: returns true if the prediction was performed, false otherwise

Reimplemented from MLBase.

Definition at line 172 of file SVM.cpp.

bool SVM::saveModelToFile ( std::fstream & file ) const

virtual

This saves the trained SVM model to a file. This overrides the saveModelToFile function in the Classifier base class.

Parameters

file	a reference to the file the SVM model will be saved to

Returns: returns true if the model was saved successfully, false otherwise

Reimplemented from MLBase.

Definition at line 499 of file SVM.cpp.

bool SVM::setC ( const Float C )

Sets the SVM C parameter. This is only used if the SVM type parameter is set to C_SVC, EPSILON_SVR or NU_SVR.

Parameters

C	the new C value

Returns: returns true if the C parameter was updated, false otherwise

Definition at line 1097 of file SVM.cpp.

bool SVM::setCoef0 ( const Float coef0 )

Sets the SVM coef0 parameter. This is only used if the SVM kernel type parameter is set to POLY_KERNEL or SIGMOID_KERNEL.

Parameters

coef0 the new coef0 value

Returns: returns true if the gamcoef0ma parameter was updated, false otherwise

Definition at line 1092 of file SVM.cpp.

bool SVM::setDegree ( const UINT degree )

Sets the SVM degree parameter. This is only used if the SVM kernel parameter is set to POLY_KERNEL.

Parameters

degree the new degree value

Returns: returns true if the degree parameter was updated, false otherwise

Definition at line 1082 of file SVM.cpp.

bool SVM::setGamma ( const Float gamma )

Sets the SVM gamma parameter. The user should only try and set this value manually if the useAutoGamma parameter has been set to false.

Parameters

gamma the new gamma value

Returns: returns true if the gamma parameter was updated, false otherwise

Definition at line 1073 of file SVM.cpp.

bool SVM::setKernelType ( const UINT kernelType )

Sets the kernel type. This should be one of the SVMKernelTypes enumeration types.

Parameters

kernelType the new kernel, options are LINEAR_KERNEL, POLY_KERNEL, RBF_KERNEL, SIGMOID_KERNEL, PRECOMPUTED_KERNEL

Returns: returns true if the kernel type was set correctly, false otherwise

Definition at line 1064 of file SVM.cpp.

bool SVM::setKFoldCrossValidationValue ( const UINT kFoldValue )

Sets the kFold cross validation value.

Parameters

kFoldValue the new kFoldValue, must be greater than 0 return returns true if the kFoldValue was set, false otherwise

Definition at line 1102 of file SVM.cpp.

bool SVM::setNu ( const Float nu )

Sets the SVM nu parameter. This is only used if the SVM type parameter is set to NU_SVC, ONE_CLASS, or NU_SVR.

Parameters

nu	the new nu value

Returns: returns true if the nu parameter was updated, false otherwise

Definition at line 1087 of file SVM.cpp.

bool SVM::setSVMType ( const UINT svmType )

Sets the SVM type. This should be one of the SVMTypes enumeration types.

Parameters

svmType the new SVM type, options are C_SVC, NU_SVC, ONE_CLASS, EPSILON_SVR, NU_SVR

Returns: returns true if the SVM type was set correctly, false otherwise

Definition at line 1056 of file SVM.cpp.

bool SVM::train_ ( ClassificationData & trainingData )

virtual

This trains the SVM model, using the labelled classification data. This overrides the train function in the Classifier base class.

Parameters

trainingData a reference to the training data

Returns: returns true if the SVM model was trained, false otherwise

Reimplemented from MLBase.

Definition at line 143 of file SVM.cpp.

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

Public Types

Public Member Functions

Protected Member Functions

Protected Attributes

Static Protected Attributes

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation