ml::AbstractLogisticRegression Class Reference

Abstract implementation, sharing the common parameters and stopping criterion. More...

#include <LogisticRegression.hpp>

Inheritance diagram for ml::AbstractLogisticRegression:

Collaboration diagram for ml::AbstractLogisticRegression:

Public Member Functions
	AbstractLogisticRegression ()
	Constructor setting default parameter values.
double	lam () const
	Returns the regularisation parameter. More...
double	absolute_tolerance () const
	Returns absolute tolerance for fitted weights.
double	relative_tolerance () const
	Returns relative tolerance for fitted weights.
unsigned int	maximum_steps () const
	Returns maximum number of steps allowed.
void	set_lam (double lam)
	Sets the regularisation parameter. More...
void	set_absolute_tolerance (double absolute_tolerance)
	Sets absolute tolerance for weight convergence. More...
void	set_relative_tolerance (double relative_tolerance)
	Sets relative tolerance for weight convergence. More...
void	set_maximum_steps (unsigned int maximum_steps)
	Sets maximum number of steps.
Public Member Functions inherited from ml::LogisticRegression
virtual	~LogisticRegression ()
	Virtual destructor.
virtual Result	fit (Eigen::Ref< const Eigen::MatrixXd > X, Eigen::Ref< const Eigen::VectorXd > y) const =0
	Fits the model and returns the result. More...

Protected Member Functions
bool	weights_converged (Eigen::Ref< const Eigen::VectorXd > old_weights, Eigen::Ref< const Eigen::VectorXd > new_weights) const
	Check if weight fitting converged. More...

Additional Inherited Members
Static Public Member Functions inherited from ml::LogisticRegression
static double	probability (Eigen::Ref< const Eigen::VectorXd > x, double y, Eigen::Ref< const Eigen::VectorXd > w)
	Calculates the probability of label given model weights and feature vector. More...
static double	log_likelihood (Eigen::Ref< const Eigen::MatrixXd > X, Eigen::Ref< const Eigen::VectorXd > y, Eigen::Ref< const Eigen::VectorXd > w, double lam)
	Calculates the posterior log-likelihood of data given model weights. More...
static void	grad_log_likelihood (Eigen::Ref< const Eigen::MatrixXd > X, Eigen::Ref< const Eigen::VectorXd > y, Eigen::Ref< const Eigen::VectorXd > w, double lam, Eigen::Ref< Eigen::VectorXd > g)
	Calculates the gradient of the posterior log-likelihood of data given model weights, over those weights. More...
static void	hessian_log_likelihood (Eigen::Ref< const Eigen::MatrixXd > X, Eigen::Ref< const Eigen::VectorXd > y, Eigen::Ref< const Eigen::VectorXd > w, double lam, Eigen::Ref< Eigen::MatrixXd > H)
	Calculates the Hessian of the posterior log-likelihood of data given model weights, over those weights. More...

Detailed Description

Abstract implementation, sharing the common parameters and stopping criterion.

The stopping criterion is ||old_weights - new_weights||_2 <= absolute_tolerance + relative_tolerance * max(||old_weights||_2, ||new_weights||_2).

Member Function Documentation

lam()

double ml::AbstractLogisticRegression::lam ( ) const

inline

Returns the regularisation parameter.

Returns

Inverse variance of the Gaussian prior for w. Cannot be negative. Set it to 0 if you want to perform maximum likelihood estimation of w.

set_lam()

void ml::AbstractLogisticRegression::set_lam

(

double

lam

)

Sets the regularisation parameter.

Parameters

lam	Inverse variance of the Gaussian prior for w. Cannot be negative. Set it to 0 if you want to perform maximum likelihood estimation of w.

Exceptions

std::domain_error

If lam is negative.

set_absolute_tolerance()

void ml::AbstractLogisticRegression::set_absolute_tolerance

(

double

absolute_tolerance

)

Sets absolute tolerance for weight convergence.

Parameters

absolute_tolerance

Cannot be negative.

Exceptions

std::domain_error

If negative.

set_relative_tolerance()

void ml::AbstractLogisticRegression::set_relative_tolerance

(

double

relative_tolerance

)

Sets relative tolerance for weight convergence.

Parameters

relative_tolerance

Cannot be negative.

Exceptions

std::domain_error

If negative.

weights_converged()

bool ml::AbstractLogisticRegression::weights_converged ( Eigen::Ref< const Eigen::VectorXd >  old_weights, Eigen::Ref< const Eigen::VectorXd >  new_weights  ) const

protected

Check if weight fitting converged.

Parameters

old_weights	Previous weight vector.
new_weights	New weight vector.

Returns

True if converged, false otherwise.

---

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

• ML/LogisticRegression.hpp