simbad.piconode.core.operator
Class ActivationFunctionWithDerivative

java.lang.Object
  extended by simbad.piconode.core.operator.Operator
      extended by simbad.piconode.core.operator.ActivationFunction
          extended by simbad.piconode.core.operator.ActivationFunctionWithDerivative
All Implemented Interfaces:
java.io.Serializable
Direct Known Subclasses:
ActivationFunction_hyperbolicTangent, ActivationFunction_logisticSigmoid, ActivationFunction_piecewiseLinear, ActivationFunction_threshold

public abstract class ActivationFunctionWithDerivative
extends ActivationFunction
implements java.io.Serializable

This class defines the prototype for any activation function, usually used in the context of artificial neural networks. This class embeds also a prototype for defining the derivative of the function needed by learning algorithm such as back-propagation.

See Also:
Serialized Form

Constructor Summary
ActivationFunctionWithDerivative()
           
 
Method Summary
abstract  double applyDerivativeFunction(double __value)
           
abstract  double applyFunction(double __value)
           
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Constructor Detail

ActivationFunctionWithDerivative

public ActivationFunctionWithDerivative()
Method Detail

applyFunction

public abstract double applyFunction(double __value)
Specified by:
applyFunction in class ActivationFunction

applyDerivativeFunction

public abstract double applyDerivativeFunction(double __value)