LinearClassifier Class Reference

The LinearClassifier class is a multi class classifier model suited for linear discriminant analysis. More...

#include <LinearModel.h>

Inheritance diagram for LinearClassifier:

Public Member Functions
	LinearClassifier (int dimension, int classes)
	~LinearClassifier ()
void	setParameter (unsigned int index, double value)
	Modifies a specific model parameter.
void	model (const Array< double > &input, Array< double > &output)
	Returns the model's answer output on the stimulus input.
int	getNumberOfClasses () const
Protected Attributes
int	numberOfClasses
Array< double >	mean
Array2D< double >	covariance
Array2D< double >	inverse
bool	bNeedsUpdate

Detailed Description

The LinearClassifier class is a multi class classifier model suited for linear discriminant analysis.

For c classes $0, \dots, c-1$ the model holds class mean vectors $ m_c $ and a shared data scatter matrix $ M $ . It predicts the class of a vector x according to the rule $\textrm{argmin}_{c} (x - m_c)^T M^{-1} (x - m_c)$ . The output is a unit vector in $ R^c $ composed of zeros and a single 1 entry in the component corresponding to the predicted class.

Definition at line 126 of file LinearModel.h.

Constructor & Destructor Documentation

LinearClassifier::LinearClassifier	(	int	dimension,
		int	classes
	)

Definition at line 232 of file LinearModel.cpp.

References bNeedsUpdate, covariance, Model::inputDimension, inverse, mean, numberOfClasses, Model::outputDimension, and Model::parameter.

LinearClassifier::~LinearClassifier ( )

Definition at line 251 of file LinearModel.cpp.

Member Function Documentation

int LinearClassifier::getNumberOfClasses ( ) const [inline]

Definition at line 135 of file LinearModel.h.

References numberOfClasses.

Referenced by LDA::optimize().

void LinearClassifier::model	(	const Array< double > &	input,
		Array< double > &	output
	)			`[virtual]`

Returns the model's answer output on the stimulus input.

This method calculates the output of the model depending on the input. The arrays input and output can either be one- or two-dimensional, depending on whether one or many patterns should be processed. The number of elements in the last dimension of these arrays must fit the inputDimension and outputDimension, in case of two-dimensional input, the number of elements in the first dimension equals the number of patterns. The method is pure virtual as it has to be implemented by the different models.

Parameters: