Support Vector Machine for more than two classes. More...

Public Member Functions
	SvmMultiClass (void)
	Constructor.

double	C (void) const
	C-parameter. More...

SvmMultiClass *	make_classifier (void) const
	Create an untrained copy of SvmMultiClass. More...

unsigned long int	max_epochs (void) const

void	max_epochs (unsigned long int)
	set maximal number of epochs in training

void	predict (const KernelLookup &input, utility::Matrix &predict) const
	Generate prediction predict from input. More...

void	set_C (double)
	set the C-Parameter

void	train (const KernelLookup &kernel, const Target &target)
	Fit paramaters based on training data. More...

Detailed Description

Support Vector Machine for more than two classes.

Member Function Documentation

double theplu::yat::classifier::SvmMultiClass::C ( void ) const

C-parameter.

SvmMultiClass* theplu::yat::classifier::SvmMultiClass::make_classifier ( void ) const

Create an untrained copy of SvmMultiClass.

Returns: A dynamically allocated SVM, which has to be deleted by the caller to avoid memory leaks.

unsigned long int theplu::yat::classifier::SvmMultiClass::max_epochs ( void ) const

Default is max_epochs set to 100,000.

void theplu::yat::classifier::SvmMultiClass::predict	(	const KernelLookup &	input,
		utility::Matrix &	predict
	)		const

Generate prediction predict from input.

Each row in predict is calculated using a binary SVM.

Parameters

input	kernel to base prediction on. Just like in SVM each column corresponds to a sample to predict, and each row corresponds to a training sample.
predict	resulting predictions. Each column corresponds to a test sample (same as corresponding column in input). Row i corresponds to class i, in other words, a large value in row i suggests that test sample belongs to class i. If there were no training samples in class i, row i is filled with NaNs.

void theplu::yat::classifier::SvmMultiClass::train	(	const KernelLookup &	kernel,
		const Target &	target
	)

Fit paramaters based on training data.

For each class samples belonging to that class are trained against the other samples using a binary SVM.

Parameters

kernel	contains the training data
target	infomration on class labels

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