slope::PGD Class Reference

Proximal Gradient Descent solver for SLOPE optimization. More...

#include <pgd.h>

Inheritance diagram for slope::PGD:

Collaboration diagram for slope::PGD:

Public Member Functions
	PGD (JitNormalization jit_normalization, bool intercept, const std::string &update_type)
	Constructs Proximal Gradient Descent solver for SLOPE optimization.
void	run (Eigen::VectorXd &beta0, Eigen::VectorXd &beta, Eigen::MatrixXd &eta, const Eigen::ArrayXd &lambda, const std::unique_ptr< Loss > &loss, const SortedL1Norm &penalty, const Eigen::VectorXd &gradient, const std::vector< int > &working_set, const Eigen::MatrixXd &x, const Eigen::VectorXd &x_centers, const Eigen::VectorXd &x_scales, const Eigen::MatrixXd &y) override
	Pure virtual function defining the solver's optimization routine.
void	run (Eigen::VectorXd &beta0, Eigen::VectorXd &beta, Eigen::MatrixXd &eta, const Eigen::ArrayXd &lambda, const std::unique_ptr< Loss > &loss, const SortedL1Norm &penalty, const Eigen::VectorXd &gradient, const std::vector< int > &working_set, const Eigen::SparseMatrix< double > &x, const Eigen::VectorXd &x_centers, const Eigen::VectorXd &x_scales, const Eigen::MatrixXd &y) override
	Pure virtual function defining the solver's optimization routine.
Public Member Functions inherited from slope::SolverBase
	SolverBase (JitNormalization jit_normalization, bool intercept)
	Constructs a base solver for SLOPE optimization.
virtual	~SolverBase ()=default
	Default desstructor.

Additional Inherited Members
Protected Attributes inherited from slope::SolverBase
JitNormalization	jit_normalization
	JIT feature normalization strategy.
bool	intercept
	If true, fits intercept term.

Detailed Description

Proximal Gradient Descent solver for SLOPE optimization.

This solver implements the proximal gradient descent algorithm with line search for solving the SLOPE optimization problem. It uses backtracking line search to automatically adjust the learning rate for optimal convergence.

Definition at line 25 of file pgd.h.

Constructor & Destructor Documentation

PGD()

slope::PGD::PGD ( JitNormalization  jit_normalization, bool  intercept, const std::string &  update_type  )

inline

Constructs Proximal Gradient Descent solver for SLOPE optimization.

Parameters

jit_normalization	Feature normalization strategy
intercept	If true, fits intercept term
update_type	Type of update strategy to use

Definition at line 34 of file pgd.h.

Member Function Documentation

run() [1/2]

void slope::PGD::run ( Eigen::VectorXd &  beta0, Eigen::VectorXd &  beta, Eigen::MatrixXd &  eta, const Eigen::ArrayXd &  lambda, const std::unique_ptr< Loss > &  loss, const SortedL1Norm &  penalty, const Eigen::VectorXd &  gradient, const std::vector< int > &  working_set, const Eigen::MatrixXd &  x, const Eigen::VectorXd &  x_centers, const Eigen::VectorXd &  x_scales, const Eigen::MatrixXd &  y  )

overridevirtual

Pure virtual function defining the solver's optimization routine.

Parameters

beta0	Intercept terms for each response
beta	Coefficients (size p x m)
eta	Linear predictor matrix (n samples x m responses)
lambda	Vector of regularization parameters
loss	Pointer to loss function object
penalty	Sorted L1 norm object for proximal operations
gradient	Gradient matrix for loss function
working_set	Vector of indices for active predictors
x	Input feature matrix (n samples x p predictors)
x_centers	Vector of feature means for centering
x_scales	Vector of feature scales for normalization
y	Response matrix (n samples x m responses)

Implements slope::SolverBase.

Definition at line 17 of file pgd.cpp.

run() [2/2]

void slope::PGD::run ( Eigen::VectorXd &  beta0, Eigen::VectorXd &  beta, Eigen::MatrixXd &  eta, const Eigen::ArrayXd &  lambda, const std::unique_ptr< Loss > &  loss, const SortedL1Norm &  penalty, const Eigen::VectorXd &  gradient, const std::vector< int > &  working_set, const Eigen::SparseMatrix< double > &  x, const Eigen::VectorXd &  x_centers, const Eigen::VectorXd &  x_scales, const Eigen::MatrixXd &  y  )

overridevirtual

Pure virtual function defining the solver's optimization routine.

Parameters

beta0	Intercept terms for each response
beta	Coefficients (size p x m)
eta	Linear predictor matrix (n samples x m responses)
lambda	Vector of regularization parameters
loss	Pointer to loss function object
penalty	Sorted L1 norm object for proximal operations
gradient	Gradient matrix for loss function
working_set	Vector of indices for active predictors
x	Input feature matrix (n samples x p predictors)
x_centers	Vector of feature means for centering
x_scales	Vector of feature scales for normalization
y	Response matrix (n samples x m responses)

Implements slope::SolverBase.

Definition at line 46 of file pgd.cpp.

---

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

• src/slope/solvers/pgd.h
• src/slope/solvers/pgd.cpp