jsat.classifiers.linear

Enum StochasticMultinomialLogisticRegression.Prior

java.lang.Object

java.lang.Enum<StochasticMultinomialLogisticRegression.Prior>

jsat.classifiers.linear.StochasticMultinomialLogisticRegression.Prior

All Implemented Interfaces:

Serializable, Comparable<StochasticMultinomialLogisticRegression.Prior>

Enclosing class:

StochasticMultinomialLogisticRegression

public static enum StochasticMultinomialLogisticRegression.Prior
extends Enum<StochasticMultinomialLogisticRegression.Prior>

Represents a prior of the coefficients that can be applied to perform regularization.

Enum Constant Summary

Enum Constants

Enum Constant and Description
CAUCHY This is a prior from the Cauchy (student-t) distribution, and it uses the extra alpha parameter.
ELASTIC This is the Elastic Net prior, and it uses the extra alpha parameter.
GAUSSIAN A Gaussian prior, this is equivalent to L2 regularization.
LAPLACE A Laplace prior, this is equivalent to L1 regularization
UNIFORM This is the Uniform prior.

Method Summary

Modifier and Type	Method and Description
protected abstract double	gradientError(double b_i, double s_i)
protected double	gradientError(double b_i, double s_i, double alpha)
protected abstract double	logProb(double b_i, double s_i)
protected double	logProb(double b_i, double s_i, double alpha)
static StochasticMultinomialLogisticRegression.Prior	valueOf(String name) Returns the enum constant of this type with the specified name.
static StochasticMultinomialLogisticRegression.Prior[]	values() Returns an array containing the constants of this enum type, in the order they are declared.

Methods inherited from class java.lang.Enum

clone, compareTo, equals, finalize, getDeclaringClass, hashCode, name, ordinal, toString, valueOf

Methods inherited from class java.lang.Object

getClass, notify, notifyAll, wait, wait, wait

Enum Constant Detail

GAUSSIAN

public static final StochasticMultinomialLogisticRegression.Prior GAUSSIAN

A Gaussian prior, this is equivalent to L₂ regularization.

LAPLACE

public static final StochasticMultinomialLogisticRegression.Prior LAPLACE

A Laplace prior, this is equivalent to L₁ regularization

ELASTIC

public static final StochasticMultinomialLogisticRegression.Prior ELASTIC

This is the Elastic Net prior, and it uses the extra alpha parameter. This prior is a mix of both LAPLACE and GAUSSIAN. Alpha should be in the range [0,1]. Alpha weight will be applied to the Laplace prior, and (1-alpha) weight will be applied to the Gaussian prior. The extreme values of this collapse into the Laplace and Gaussian priors.

CAUCHY

public static final StochasticMultinomialLogisticRegression.Prior CAUCHY

This is a prior from the Cauchy (student-t) distribution, and it uses the extra alpha parameter. Alpha should be in the range (0, Infty).

UNIFORM

public static final StochasticMultinomialLogisticRegression.Prior UNIFORM

This is the Uniform prior. The uniform prior is equivalent to no regularization.

Method Detail

values

public static StochasticMultinomialLogisticRegression.Prior[] values()

Returns an array containing the constants of this enum type, in the order they are declared. This method may be used to iterate over the constants as follows:

for (StochasticMultinomialLogisticRegression.Prior c : StochasticMultinomialLogisticRegression.Prior.values())
    System.out.println(c);

Returns:

an array containing the constants of this enum type, in the order they are declared

valueOf

public static StochasticMultinomialLogisticRegression.Prior valueOf(String name)

Returns the enum constant of this type with the specified name. The string must match exactly an identifier used to declare an enum constant in this type. (Extraneous whitespace characters are not permitted.)

Parameters:

name - the name of the enum constant to be returned.

Returns:

the enum constant with the specified name

Throws:

IllegalArgumentException - if this enum type has no constant with the specified name

NullPointerException - if the argument is null

gradientError

protected abstract double gradientError(double b_i,
                                        double s_i)

gradientError

protected double gradientError(double b_i,
                               double s_i,
                               double alpha)

logProb

protected abstract double logProb(double b_i,
                                  double s_i)

logProb

protected double logProb(double b_i,
                         double s_i,
                         double alpha)