SG++-Doxygen-Documentation
sgpp::solver::RegularizationFunction Class Referenceabstract

The RegularizationFunction class is a baseclass for regularization functions that can be used in conjunction with a proximal solver. More...

#include <RegularizationFunction.hpp>

Inheritance diagram for sgpp::solver::RegularizationFunction:
sgpp::solver::ElasticNetFunction sgpp::solver::GroupLassoFunction sgpp::solver::LassoFunction sgpp::solver::RidgeFunction sgpp::solver::ZeroFunction

Public Member Functions

virtual double eval (base::DataVector weights)=0
 eval evaluates the regularization function for weights. More...
 
virtual base::DataVector prox (const base::DataVector &weights, double stepsize)=0
 prox evaluates the proximal operator for the function for weights. More...
 

Detailed Description

The RegularizationFunction class is a baseclass for regularization functions that can be used in conjunction with a proximal solver.

Member Function Documentation

◆ eval()

virtual double sgpp::solver::RegularizationFunction::eval ( base::DataVector  weights)
pure virtual

eval evaluates the regularization function for weights.

Parameters
weights

Implemented in sgpp::solver::GroupLassoFunction, sgpp::solver::ElasticNetFunction, sgpp::solver::LassoFunction, sgpp::solver::RidgeFunction, and sgpp::solver::ZeroFunction.

◆ prox()

virtual base::DataVector sgpp::solver::RegularizationFunction::prox ( const base::DataVector weights,
double  stepsize 
)
pure virtual

prox evaluates the proximal operator for the function for weights.

Parameters
weights
stepsizeis the stepsize used for the proximal step

Implemented in sgpp::solver::GroupLassoFunction, sgpp::solver::ElasticNetFunction, sgpp::solver::LassoFunction, sgpp::solver::RidgeFunction, and sgpp::solver::ZeroFunction.


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