org.neuroph.util.random
Class DistortRandomizer

java.lang.Object
  extended by org.neuroph.util.random.WeightsRandomizer
      extended by org.neuroph.util.random.DistortRandomizer

public class DistortRandomizer
extends WeightsRandomizer

This class provides distort randomization technique, which distorts existing weight values using specified distortion factor. Weights are distorted using following formula: newWeightValue = currentWeightValue + (distortionFactor - (random * distortionFactor * 2))

Author:
Zoran Sevarac

Field Summary
 
Fields inherited from class org.neuroph.util.random.WeightsRandomizer
randomGenerator
 
Constructor Summary
DistortRandomizer(double distortionFactor)
          Create a new instance of DistortRandomizer with specified distortion factor
 
Method Summary
 void randomize(NeuralNetwork neuralNetwork)
          Iterate all layers, neurons and connection weight and apply distort randomization
 
Methods inherited from class org.neuroph.util.random.WeightsRandomizer
getRandomGenerator, nextRandomWeight
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Constructor Detail

DistortRandomizer

public DistortRandomizer(double distortionFactor)
Create a new instance of DistortRandomizer with specified distortion factor

Parameters:
distortionFactor - amount to distort existing weights
Method Detail

randomize

public void randomize(NeuralNetwork neuralNetwork)
Iterate all layers, neurons and connection weight and apply distort randomization

Overrides:
randomize in class WeightsRandomizer
Parameters:
neuralNetwork -


Copyright © 2012. All Rights Reserved.