Softmax Regression is a classifier which can be used for classification when the data available can take two or more class values. More...

Public Member Functions
	SoftmaxRegression (const size_t inputSize=0, const size_t numClasses=0, const bool fitIntercept=false)
	Initialize the SoftmaxRegression without performing training. More...

	SoftmaxRegression (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 SoftmaxRegression class with the provided data and labels. More...

	SoftmaxRegression (OptimizerType< SoftmaxRegressionFunction > &optimizer)
	Construct the softmax regression model with the given training data. More...

double	ComputeAccuracy (const arma::mat &testData, const arma::Row< size_t > &labels) const
	Computes accuracy of the learned model given the feature data and the labels associated with each data point. More...

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

bool	FitIntercept () const
	Gets the intercept term flag. We can't change this after training. More...

double &	Lambda ()
	Sets the regularization parameter. More...

double	Lambda () const
	Gets the regularization parameter. More...

size_t &	NumClasses ()
	Sets the number of classes. More...

size_t	NumClasses () const
	Gets the number of classes. More...

arma::mat &	Parameters ()
	Get the model parameters. More...

const arma::mat &	Parameters () const
	Get the model parameters. More...

void	Predict (const arma::mat &testData, arma::Row< size_t > &predictions) const
	Predict the class labels for the provided feature points. More...

template<typename Archive >
void	Serialize (Archive &ar, const unsigned int)
	Serialize the SoftmaxRegression model. More...

double	Train (OptimizerType< SoftmaxRegressionFunction > &optimizer)
	Train the softmax regression model with the given optimizer. More...

double	Train (const arma::mat &data, const arma::Row< size_t > &labels, const size_t numClasses)
	Train the softmax regression with the given training data. More...

Private Attributes
bool	fitIntercept
	Intercept term flag. More...

double	lambda
	L2-regularization constant. More...

size_t	numClasses
	Number of classes. More...

arma::mat	parameters
	Parameters after optimization. More...

Detailed Description

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>
class mlpack::regression::SoftmaxRegression< OptimizerType >

Softmax Regression is a classifier which can be used for classification when the data available can take two or more class values.

It is a generalization of Logistic Regression (which is used only for binary classification). The model has a different set of parameters for each class, but can be easily converted into a vectorized implementation as has been done in this module. The model can be used for direct classification of feature data or in conjunction with unsupervised learning methods. More technical details about the model can be found on the following webpage:

http://ufldl.stanford.edu/wiki/index.php/Softmax_Regression

An example on how to use the interface is shown below:

arma::mat train_data; // Training data matrix.
arma::vec labels; // Labels associated with the data.
const size_t inputSize = 784; // Size of input feature vector.
const size_t numClasses = 10; // Number of classes.
// Train the model using default options.
SoftmaxRegression<> regressor1(train_data, labels, inputSize, numClasses);
const size_t numBasis = 5; // Parameter required for L-BFGS algorithm.
const size_t numIterations = 100; // Maximum number of iterations.
// Use an instantiated optimizer for the training.
SoftmaxRegressionFunction srf(train_data, labels, inputSize, numClasses);
L_BFGS<SoftmaxRegressionFunction> optimizer(srf, numBasis, numIterations);
SoftmaxRegression<L_BFGS> regressor2(optimizer);
arma::mat test_data; // Test data matrix.
arma::vec predictions1, predictions2; // Vectors to store predictions in.
// Obtain predictions from both the learned models.
regressor1.Predict(test_data, predictions1);
regressor2.Predict(test_data, predictions2);

Definition at line 65 of file softmax_regression.hpp.

Constructor & Destructor Documentation

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

mlpack::regression::SoftmaxRegression< OptimizerType >::SoftmaxRegression	(	const size_t	inputSize = `0`,
		const size_t	numClasses = `0`,
		const bool	fitIntercept = `false`
	)

Initialize the SoftmaxRegression without performing training.

Default value of lambda is 0.0001. Be sure to use Train() before calling Predict() or ComputeAccuracy(), otherwise the results may be meaningless.

Parameters

inputSize	Size of the input feature vector.
numClasses	Number of classes for classification.
fitIntercept	add intercept term or not.

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

mlpack::regression::SoftmaxRegression< OptimizerType >::SoftmaxRegression	(	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 SoftmaxRegression class with the provided data and labels.

This will train the model. Optionally, the parameter 'lambda' can be passed, which controls the amount of L2-regularization in the objective function. By default, the model takes a small value.

Parameters

data	Input training features. Each column associate with one sample
labels	Labels associated with the feature data.
inputSize	Size of the input feature vector.
numClasses	Number of classes for classification.
lambda	L2-regularization constant.
fitIntercept	add intercept term or not.

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

mlpack::regression::SoftmaxRegression< OptimizerType >::SoftmaxRegression ( OptimizerType< SoftmaxRegressionFunction > & optimizer )

Construct the softmax regression model with the given training data.

This will train the model. This overload takes an already instantiated optimizer and uses it to train the model. The optimizer should hold an instantiated SoftmaxRegressionFunction object for the function to operate upon. This option should be preferred when the optimizer options are to be changed.

Parameters

optimizer Instantiated optimizer with instantiated error function.

Member Function Documentation

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

double mlpack::regression::SoftmaxRegression< OptimizerType >::ComputeAccuracy	(	const arma::mat &	testData,
		const arma::Row< size_t > &	labels
	)		const

Computes accuracy of the learned model given the feature data and the labels associated with each data point.

Predictions are made using the provided data and are compared with the actual labels.

Parameters

testData	Matrix of data points using which predictions are made.
labels	Vector of labels associated with the data.

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

size_t mlpack::regression::SoftmaxRegression< OptimizerType >::FeatureSize ( ) const

inline

Gets the features size of the training data.

Definition at line 171 of file softmax_regression.hpp.

References mlpack::regression::SoftmaxRegression< OptimizerType >::fitIntercept, and mlpack::regression::SoftmaxRegression< OptimizerType >::parameters.

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

bool mlpack::regression::SoftmaxRegression< OptimizerType >::FitIntercept ( ) const

inline

Gets the intercept term flag. We can't change this after training.

Definition at line 163 of file softmax_regression.hpp.

References mlpack::regression::SoftmaxRegression< OptimizerType >::fitIntercept.

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

double& mlpack::regression::SoftmaxRegression< OptimizerType >::Lambda ( )

inline

Sets the regularization parameter.

Definition at line 158 of file softmax_regression.hpp.

References mlpack::regression::SoftmaxRegression< OptimizerType >::lambda.

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

double mlpack::regression::SoftmaxRegression< OptimizerType >::Lambda ( ) const

inline

Gets the regularization parameter.

Definition at line 160 of file softmax_regression.hpp.

References mlpack::regression::SoftmaxRegression< OptimizerType >::lambda.

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

size_t& mlpack::regression::SoftmaxRegression< OptimizerType >::NumClasses ( )

inline

Sets the number of classes.

Definition at line 153 of file softmax_regression.hpp.

References mlpack::regression::SoftmaxRegression< OptimizerType >::numClasses.

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

size_t mlpack::regression::SoftmaxRegression< OptimizerType >::NumClasses ( ) const

inline

Gets the number of classes.

Definition at line 155 of file softmax_regression.hpp.

References mlpack::regression::SoftmaxRegression< OptimizerType >::numClasses.

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

arma::mat& mlpack::regression::SoftmaxRegression< OptimizerType >::Parameters ( )

inline

Get the model parameters.

Definition at line 166 of file softmax_regression.hpp.

References mlpack::regression::SoftmaxRegression< OptimizerType >::parameters.

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

const arma::mat& mlpack::regression::SoftmaxRegression< OptimizerType >::Parameters ( ) const

inline

Get the model parameters.

Definition at line 168 of file softmax_regression.hpp.

References mlpack::regression::SoftmaxRegression< OptimizerType >::parameters.

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

void mlpack::regression::SoftmaxRegression< OptimizerType >::Predict	(	const arma::mat &	testData,
		arma::Row< size_t > &	predictions
	)		const

Predict the class labels for the provided feature points.

The function calculates the probabilities for every class, given a data point. It then chooses the class which has the highest probability among all.

Parameters

testData	Matrix of data points for which predictions are to be made.
predictions	Vector to store the predictions in.

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

template<typename Archive >

void mlpack::regression::SoftmaxRegression< OptimizerType >::Serialize	(	Archive &	ar,
		const unsigned	int
	)

inline

Serialize the SoftmaxRegression model.

Definition at line 179 of file softmax_regression.hpp.

References mlpack::data::CreateNVP(), mlpack::regression::SoftmaxRegression< OptimizerType >::fitIntercept, mlpack::regression::SoftmaxRegression< OptimizerType >::lambda, and mlpack::regression::SoftmaxRegression< OptimizerType >::parameters.

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

double mlpack::regression::SoftmaxRegression< OptimizerType >::Train ( OptimizerType< SoftmaxRegressionFunction > & optimizer )

Train the softmax regression model with the given optimizer.

The optimizer should hold an instantiated SoftmaxRegressionFunction object for the function to operate upon. This option should be preferred when the optimizer options are to be changed.

Parameters

optimizer Instantiated optimizer with instantiated error function.

Returns: Objective value of the final point.

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

double mlpack::regression::SoftmaxRegression< OptimizerType >::Train	(	const arma::mat &	data,
		const arma::Row< size_t > &	labels,
		const size_t	numClasses
	)

Train the softmax regression with the given training data.

Parameters

data	Input data with each column as one example.
labels	Labels associated with the feature data.
numClasses	Number of classes for classification.

Returns: Objective value of the final point.

Member Data Documentation

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

bool mlpack::regression::SoftmaxRegression< OptimizerType >::fitIntercept

private

Intercept term flag.

Definition at line 197 of file softmax_regression.hpp.

Referenced by mlpack::regression::SoftmaxRegression< OptimizerType >::FeatureSize(), mlpack::regression::SoftmaxRegression< OptimizerType >::FitIntercept(), and mlpack::regression::SoftmaxRegression< OptimizerType >::Serialize().

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

double mlpack::regression::SoftmaxRegression< OptimizerType >::lambda

private

L2-regularization constant.

Definition at line 195 of file softmax_regression.hpp.

Referenced by mlpack::regression::SoftmaxRegression< OptimizerType >::Lambda(), and mlpack::regression::SoftmaxRegression< OptimizerType >::Serialize().

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

size_t mlpack::regression::SoftmaxRegression< OptimizerType >::numClasses

private

Number of classes.

Definition at line 193 of file softmax_regression.hpp.

Referenced by mlpack::regression::SoftmaxRegression< OptimizerType >::NumClasses().

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>

arma::mat mlpack::regression::SoftmaxRegression< OptimizerType >::parameters

private

Parameters after optimization.

Definition at line 191 of file softmax_regression.hpp.

Referenced by mlpack::regression::SoftmaxRegression< OptimizerType >::FeatureSize(), mlpack::regression::SoftmaxRegression< OptimizerType >::Parameters(), and mlpack::regression::SoftmaxRegression< OptimizerType >::Serialize().

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

src/mlpack/methods/softmax_regression/softmax_regression.hpp

Public Member Functions

Private Attributes

Detailed Description

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS> class mlpack::regression::SoftmaxRegression< OptimizerType >

Constructor & Destructor Documentation

Member Function Documentation

Member Data Documentation

template<template< typename > class OptimizerType = mlpack::optimization::L_BFGS>
class mlpack::regression::SoftmaxRegression< OptimizerType >