hex.glm

Class GLMModel.GLMWeightsFun

java.lang.Object

water.Iced

hex.glm.GLMModel.GLMWeightsFun

All Implemented Interfaces:

java.io.Externalizable, java.io.Serializable, java.lang.Cloneable, water.Freezable

Enclosing class:

GLMModel

public static class GLMModel.GLMWeightsFun
extends water.Iced

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
GLMModel.GLMParameters.Family	_family

Constructor Summary

Constructors

Constructor and Description
GLMWeightsFun(GLMModel.GLMParameters.Family fam, GLMModel.GLMParameters.Link link, double var_power, double link_power, double theta, double dispersion, boolean varPowerEstimation)
GLMWeightsFun(GLMModel.GLMParameters parms)

Method Summary

Modifier and Type	Method and Description
GLMModel.GLMWeights	computeWeights(double y, double eta, double off, double w, GLMModel.GLMWeights x)
double	deviance(double yr, double ym)
double	deviance(float yr, float ym)
double	likelihood(double yr, double ym)
double	likelihood(double w, double yr, double ym)
void	likelihoodAndDeviance(double yr, GLMModel.GLMWeights x, double w)
double	link(double x) Given the estimated model output x, we want to find the linear part which is transpose(beta)*p+intercept if beta does not contain the intercept.
double	linkDeriv(double x)
double	linkInv(double x) Given the linear combination transpose(beta)*p+intercept (if beta does not contain the intercept), this method will provide the estimated model output.
double	linkInvDeriv(double x)
double	linkInvDeriv2(double x)
double	variance(double mu)

Methods inherited from class water.Iced

asBytes, clone, copyOver, frozenType, read, readExternal, readJSON, reloadFromBytes, toJsonBytes, toJsonString, write, writeExternal, writeJSON

Methods inherited from class java.lang.Object

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

Field Detail

_family

public final GLMModel.GLMParameters.Family _family

Constructor Detail

GLMWeightsFun

public GLMWeightsFun(GLMModel.GLMParameters parms)

GLMWeightsFun

public GLMWeightsFun(GLMModel.GLMParameters.Family fam,
                     GLMModel.GLMParameters.Link link,
                     double var_power,
                     double link_power,
                     double theta,
                     double dispersion,
                     boolean varPowerEstimation)

Method Detail

link

public final double link(double x)

Given the estimated model output x, we want to find the linear part which is transpose(beta)*p+intercept if beta does not contain the intercept.

linkInvDeriv

public final double linkInvDeriv(double x)

linkInvDeriv2

public final double linkInvDeriv2(double x)

linkDeriv

public final double linkDeriv(double x)

linkInv

public final double linkInv(double x)

Given the linear combination transpose(beta)*p+intercept (if beta does not contain the intercept), this method will provide the estimated model output.

variance

public final double variance(double mu)

deviance

public final double deviance(double yr,
                             double ym)

deviance

public final double deviance(float yr,
                             float ym)

likelihoodAndDeviance

public final void likelihoodAndDeviance(double yr,
                                        GLMModel.GLMWeights x,
                                        double w)

likelihood

public final double likelihood(double w,
                               double yr,
                               double ym)

likelihood

public final double likelihood(double yr,
                               double ym)

computeWeights

public GLMModel.GLMWeights computeWeights(double y,
                                          double eta,
                                          double off,
                                          double w,
                                          GLMModel.GLMWeights x)