Public Member Functions
	SoftmaxRegressionFunction (const arma::mat &data, const arma::Row< size_t > &labels, const size_t numClasses, const double lambda=0.0001, const bool fitIntercept=false)
	Construct the Softmax Regression objective function with the given parameters. More...

const arma::mat	InitializeWeights ()
	Initializes the parameters of the model to suitable values. More...

void	Shuffle ()
	Shuffle the dataset. More...

void	GetGroundTruthMatrix (const arma::Row< size_t > &labels, arma::sp_mat &groundTruth)
	Constructs the ground truth label matrix with the passed labels. More...

void	GetProbabilitiesMatrix (const arma::mat &parameters, arma::mat &probabilities, const size_t start, const size_t batchSize) const
	Evaluate the probabilities matrix with the passed parameters. More...

double	Evaluate (const arma::mat &parameters) const
	Evaluates the objective function of the softmax regression model using the given parameters. More...

double	Evaluate (const arma::mat &parameters, const size_t start, const size_t batchSize=1) const
	Evaluate the objective function of the softmax regression model for a subset of the data points using the given parameters. More...

void	Gradient (const arma::mat &parameters, arma::mat &gradient) const
	Evaluates the gradient values of the objective function given the current set of parameters. More...

void	Gradient (const arma::mat &parameters, const size_t start, arma::mat &gradient, const size_t batchSize=1) const
	Evaluate the gradient of the objective function given the current set of parameters, on a subset of the data. More...

void	PartialGradient (const arma::mat &parameters, size_t j, arma::sp_mat &gradient) const
	Evaluates the gradient values of the objective function given the current set of parameters for a single feature indexed by j. More...

const arma::mat &	GetInitialPoint () const
	Return the initial point for the optimization.

size_t	NumClasses () const
	Gets the number of classes.

size_t	NumFeatures () const
	Gets the features size of the training data.

size_t	NumFunctions () const
	Return the number of separable functions (the number of predictor points).

double &	Lambda ()
	Sets the regularization parameter.

double	Lambda () const
	Gets the regularization parameter.

bool	FitIntercept () const
	Gets the intercept flag.

Static Public Member Functions
static const arma::mat	InitializeWeights (const size_t featureSize, const size_t numClasses, const bool fitIntercept=false)
	Initialize Softmax Regression weights (trainable parameters) with the given parameters. More...

static void	InitializeWeights (arma::mat &weights, const size_t featureSize, const size_t numClasses, const bool fitIntercept=false)
	Initialize Softmax Regression weights (trainable parameters) with the given parameters. More...

Constructor & Destructor Documentation

◆ SoftmaxRegressionFunction()

SoftmaxRegressionFunction::SoftmaxRegressionFunction	(	const arma::mat &	data,
		const arma::Row< size_t > &	labels,
		const size_t	numClasses,
		const double	lambda = `0.0001`,
		const bool	fitIntercept = `false`
	)

Construct the Softmax Regression objective function with the given parameters.

Parameters

data	Input training data, each column associate with one sample
labels	Labels associated with the feature data.
numClasses	Number of classes for classification.
lambda	L2-regularization constant.
fitIntercept	Intercept term flag.

Member Function Documentation

◆ Evaluate() [1/2]

double SoftmaxRegressionFunction::Evaluate ( const arma::mat & parameters ) const

Evaluates the objective function of the softmax regression model using the given parameters.

Evaluates the objective function given the parameters.

The cost function has terms for the log likelihood error and the regularization cost. The objective function takes a low value when the model generalizes well for the given training data, while having small parameter values.

Parameters

parameters Current values of the model parameters.

◆ Evaluate() [2/2]

double SoftmaxRegressionFunction::Evaluate	(	const arma::mat &	parameters,
		const size_t	start,
		const size_t	batchSize = `1`
	)		const

Evaluate the objective function of the softmax regression model for a subset of the data points using the given parameters.

Evaluate the objective function for the given points given the parameters.

The cost function has terms for the log likelihood error and the regularization cost. The objective function takes a low value when the model generalizes well for the given training data, while having small parameter values.

Parameters

parameters	Current values of the model parameters.
start	First index of the data points to use.
batchSize	Number of data points to evaluate objective for.

◆ GetGroundTruthMatrix()

void SoftmaxRegressionFunction::GetGroundTruthMatrix	(	const arma::Row< size_t > &	labels,
		arma::sp_mat &	groundTruth
	)

Constructs the ground truth label matrix with the passed labels.

This is equivalent to applying the indicator function to the training labels.

Parameters

labels	Labels associated with the training data.
groundTruth	Pointer to arma::mat which stores the computed matrix.

The output is in the form of a matrix, which leads to simpler calculations in the Evaluate() and Gradient() methods.

◆ GetProbabilitiesMatrix()

void SoftmaxRegressionFunction::GetProbabilitiesMatrix	(	const arma::mat &	parameters,
		arma::mat &	probabilities,
		const size_t	start,
		const size_t	batchSize
	)		const

Evaluate the probabilities matrix with the passed parameters.

Evaluate the probabilities matrix.

probabilities(i, j) = \( exp(\theta_i * data_j) / sum_k(exp(\theta_k * data_j)) \). It represents the probability of data_j belongs to class i.

Parameters

parameters	Current values of the model parameters.
probabilities	Pointer to arma::mat which stores the probabilities.
start	Index of point to start at.
batchSize	Number of points to calculate probabilities for.

If fitIntercept flag is true, it should consider the parameters.cols(0) intercept term.

◆ Gradient() [1/2]

void SoftmaxRegressionFunction::Gradient	(	const arma::mat &	parameters,
		arma::mat &	gradient
	)		const

Evaluates the gradient values of the objective function given the current set of parameters.

Calculates and stores the gradient values given a set of parameters.

The function calculates the probabilities for each class given the parameters, and computes the gradients based on the difference from the ground truth.

Parameters

parameters	Current values of the model parameters.
gradient	Matrix where gradient values will be stored.

◆ Gradient() [2/2]

void SoftmaxRegressionFunction::Gradient	(	const arma::mat &	parameters,
		const size_t	start,
		arma::mat &	gradient,
		const size_t	batchSize = `1`
	)		const

Evaluate the gradient of the objective function given the current set of parameters, on a subset of the data.

The function calculates the probabilities for each class given the parameters, and computes the gradients based on the difference from the ground truth.

Parameters

parameters	Current values of the model parameters.
start	First index of the data points to use.
gradient	Matrix to store gradient into.
batchSize	Number of data points to evaluate gradient for.

◆ InitializeWeights() [1/3]

const arma::mat SoftmaxRegressionFunction::InitializeWeights ( )

Initializes the parameters of the model to suitable values.

Initializes parameter weights to random values taken from a scaled standard normal distribution.

The weights cannot be initialized to zero, as that will lead to each class output being the same.

◆ InitializeWeights() [2/3]

const arma::mat SoftmaxRegressionFunction::InitializeWeights	(	const size_t	featureSize,
		const size_t	numClasses,
		const bool	fitIntercept = `false`
	)

static

Initialize Softmax Regression weights (trainable parameters) with the given parameters.

Parameters

featureSize	The number of features in the training set.
numClasses	Number of classes for classification.
fitIntercept	If true, an intercept is fitted.

Returns: Initialized model weights.

◆ InitializeWeights() [3/3]

void SoftmaxRegressionFunction::InitializeWeights	(	arma::mat &	weights,
		const size_t	featureSize,
		const size_t	numClasses,
		const bool	fitIntercept = `false`
	)

static

Initialize Softmax Regression weights (trainable parameters) with the given parameters.

Parameters

weights	This will be filled with the initialized model weights.
featureSize	The number of features in the training set.
numClasses	Number of classes for classification.
fitIntercept	Intercept term flag.

◆ PartialGradient()

void SoftmaxRegressionFunction::PartialGradient	(	const arma::mat &	parameters,
		size_t	j,
		arma::sp_mat &	gradient
	)		const

Evaluates the gradient values of the objective function given the current set of parameters for a single feature indexed by j.

Parameters

parameters	Current values of the model parameters.
j	The index of the feature with respect to which the partial gradient is to be computed.
gradient	Out param for the gradient value.

◆ Shuffle()

void SoftmaxRegressionFunction::Shuffle ( )

Shuffle the dataset.

Shuffle the data.

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

src/mlpack/methods/softmax_regression/softmax_regression_function.hpp
src/mlpack/methods/softmax_regression/softmax_regression_function.cpp

Public Member Functions

Static Public Member Functions

Constructor & Destructor Documentation

◆ SoftmaxRegressionFunction()

Member Function Documentation

◆ Evaluate() [1/2]

◆ Evaluate() [2/2]

◆ GetGroundTruthMatrix()

◆ GetProbabilitiesMatrix()

◆ Gradient() [1/2]

◆ Gradient() [2/2]

◆ InitializeWeights() [1/3]

◆ InitializeWeights() [2/3]

◆ InitializeWeights() [3/3]

◆ PartialGradient()

◆ Shuffle()