org.apache.commons.math3.ml.neuralnet.sofm.util

Class QuasiSigmoidDecayFunction

java.lang.Object

org.apache.commons.math3.ml.neuralnet.sofm.util.QuasiSigmoidDecayFunction

public class QuasiSigmoidDecayFunction
extends java.lang.Object

Decay function whose shape is similar to a sigmoid.
Class is immutable.

Since:

3.3

Field Summary

Fields

Modifier and Type	Field and Description
private double	scale See value(long).
private Logistic	sigmoid Sigmoid.

Constructor Summary

Constructors

Constructor and Description
QuasiSigmoidDecayFunction(double initValue, double slope, long numCall) Creates an instance.

Method Summary

Methods

Modifier and Type	Method and Description
double	value(long numCall) Computes the value of the learning factor.

Methods inherited from class java.lang.Object

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

Field Detail

sigmoid

private final Logistic sigmoid

Sigmoid.

scale

private final double scale

See value(long).

Constructor Detail

QuasiSigmoidDecayFunction

public QuasiSigmoidDecayFunction(double initValue,
                         double slope,
                         long numCall)

Creates an instance. The function f will have the following properties:

• f(0) = initValue
• numCall is the inflexion point
• slope = f'(numCall)

Parameters:

initValue - Initial value, i.e. value(0).

slope - Value of the function derivative at numCall.

numCall - Inflexion point.

Throws:

NotStrictlyPositiveException - if initValue <= 0.

NumberIsTooLargeException - if slope >= 0.

NotStrictlyPositiveException - if numCall <= 0.

Method Detail

value

public double value(long numCall)

Computes the value of the learning factor.

Parameters:

numCall - Current step of the training task.

Returns:

the value of the function at numCall.