hex.glm

Class GLM.ProximalGradientSolver

java.lang.Object

hex.optimization.L_BFGS.GradientSolver

hex.glm.GLM.ProximalGradientSolver

Enclosing class:

GLM

public static class GLM.ProximalGradientSolver
extends L_BFGS.GradientSolver

Simple wrapper around gradient computation, adding proximal penalty

Constructor Summary

Constructors

Constructor and Description
GLM.ProximalGradientSolver(L_BFGS.GradientSolver s, double[] betaGiven, double[] rho)

Method Summary

Methods

Modifier and Type	Method and Description
L_BFGS.GradientInfo	getGradient(double[] beta) Evaluate gradient at solution beta.
double[]	getObjVals(double[] beta, double[] pk, int nSteps, double initialStep, double stepDec) Evaluate objective values at k line search points beta_k.

Methods inherited from class hex.optimization.L_BFGS.GradientSolver

doLineSearch

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

GLM.ProximalGradientSolver

public GLM.ProximalGradientSolver(L_BFGS.GradientSolver s,
                          double[] betaGiven,
                          double[] rho)

Method Detail

getGradient

public L_BFGS.GradientInfo getGradient(double[] beta)

Description copied from class: L_BFGS.GradientSolver

Evaluate gradient at solution beta.

Specified by:

getGradient in class L_BFGS.GradientSolver

Returns:

getObjVals

public double[] getObjVals(double[] beta,
                  double[] pk,
                  int nSteps,
                  double initialStep,
                  double stepDec)

Description copied from class: L_BFGS.GradientSolver

Evaluate objective values at k line search points beta_k. When used as part of default line search behavior, the line search points are expected to be beta_k = beta + direction * _startStep * _stepDec^k

Specified by:

getObjVals in class L_BFGS.GradientSolver

Parameters:

beta - - initial vector of coefficients

pk - - search direction

Returns:

objective values evaluated at k line-search points beta + pk*step[k]