com.rapidminer.operator.learner.functions.neuralnet
Class ActivationFunction

java.lang.Object
  com.rapidminer.operator.learner.functions.neuralnet.ActivationFunction

Direct Known Subclasses:

LinearFunction, SigmoidFunction

public abstract class ActivationFunction
extends java.lang.Object

This is the activation function of a neural net node. This class performs the calculation of the node's output values as well as the error calculation and the update of the weights.

Author:

Ingo Mierswa

Constructor Summary

ActivationFunction()

Method Summary

abstract double	calculateError(InnerNode node, Example example) Calculates the error.
abstract double	calculateValue(InnerNode node, Example example) Calculates the value.
abstract java.lang.String	getTypeName() Delivers the name of this activation function.
void	update(InnerNode node, Example example, double learningRate, double momentum) Calculates the update of the weights.

Methods inherited from class java.lang.Object

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

Constructor Detail

ActivationFunction

public ActivationFunction()

Method Detail

getTypeName

public abstract java.lang.String getTypeName()

Delivers the name of this activation function.

calculateValue

public abstract double calculateValue(InnerNode node,
                                      Example example)

Calculates the value.

calculateError

public abstract double calculateError(InnerNode node,
                                      Example example)

Calculates the error.

update

public void update(InnerNode node,
                   Example example,
                   double learningRate,
                   double momentum)

Calculates the update of the weights.