simbad.piconode.toolbox
Class Tools

java.lang.Object
  extended by simbad.piconode.toolbox.Tools

public class Tools
extends java.lang.Object


Field Summary
private static boolean DEBUG_MODE
           
 
Constructor Summary
Tools()
           
 
Method Summary
static double getArcWeightRandomInitValue()
          Used to randomly initialise the weight of arc.
static double getArcWeightRandomInitValue(double __offset, double __range)
          Used to randomly initialise the weight of the arc.
static void initializeRandomArcWeights(FeedForwardNeuralNetworkForBackPropLearning __network)
          initialize ALL arcs weights includind those related to the bias neuron (if any).
static void initializeRandomArcWeights(FeedForwardNeuralNetworkForBackPropLearning __network, double __offset, double __range)
          initialize ALL arcs weights includind those related to the bias neuron (if any).
private static void initializeRandomArcWeightsFromBiasNeuron(FeedForwardNeuralNetworkForBackPropLearning __network)
          initialize only arcs weights that are connected to the bias neuron.
private static void initializeRandomArcWeightsFromBiasNeuron(FeedForwardNeuralNetworkForBackPropLearning __network, double __offset, double __range)
          initialize only arcs weights that are connected to the bias neuron.
private static void initializeRandomArcWeightsFromRegisteredNeurons(FeedForwardNeuralNetworkForBackPropLearning __network)
          initialize all arcs weights, except for arcs connected to the bias neuron (if any).
private static void initializeRandomArcWeightsFromRegisteredNeurons(FeedForwardNeuralNetworkForBackPropLearning __network, double __offset, double __range)
          initialize all arcs weights, except for arcs connected to the bias neuron (if any).
static double tanh(double __value)
          apply the hyperbolic tangent function.
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Field Detail

DEBUG_MODE

private static boolean DEBUG_MODE
Constructor Detail

Tools

public Tools()
Method Detail

getArcWeightRandomInitValue

public static double getArcWeightRandomInitValue()
Used to randomly initialise the weight of arc. Return a value between [ -1 , +1 ].


getArcWeightRandomInitValue

public static double getArcWeightRandomInitValue(double __offset,
                                                 double __range)
Used to randomly initialise the weight of the arc. Return a value within [ __offset , __offset + __range ] (offset value can be negative).

Parameters:
__offset -
__range -
Returns:

initializeRandomArcWeights

public static void initializeRandomArcWeights(FeedForwardNeuralNetworkForBackPropLearning __network)
initialize ALL arcs weights includind those related to the bias neuron (if any). Initialisation value within [ -1 , +1 ]

Parameters:
__network -

initializeRandomArcWeightsFromRegisteredNeurons

private static void initializeRandomArcWeightsFromRegisteredNeurons(FeedForwardNeuralNetworkForBackPropLearning __network)
initialize all arcs weights, except for arcs connected to the bias neuron (if any). Initialisation value within [ -1 , +1 ]

Parameters:
__network -

initializeRandomArcWeightsFromBiasNeuron

private static void initializeRandomArcWeightsFromBiasNeuron(FeedForwardNeuralNetworkForBackPropLearning __network)
initialize only arcs weights that are connected to the bias neuron. Initialisation value within [ -1 , +1 ]

Parameters:
__network -

initializeRandomArcWeights

public static void initializeRandomArcWeights(FeedForwardNeuralNetworkForBackPropLearning __network,
                                              double __offset,
                                              double __range)
initialize ALL arcs weights includind those related to the bias neuron (if any). Initialisation value within [ __offset , __offset + __range ] (offset value can be negative).

Parameters:
__network -

initializeRandomArcWeightsFromRegisteredNeurons

private static void initializeRandomArcWeightsFromRegisteredNeurons(FeedForwardNeuralNetworkForBackPropLearning __network,
                                                                    double __offset,
                                                                    double __range)
initialize all arcs weights, except for arcs connected to the bias neuron (if any). Initialisation value within [ __offset , __offset + __range ] (offset value can be negative).

Parameters:
__network -

initializeRandomArcWeightsFromBiasNeuron

private static void initializeRandomArcWeightsFromBiasNeuron(FeedForwardNeuralNetworkForBackPropLearning __network,
                                                             double __offset,
                                                             double __range)
initialize only arcs weights that are connected to the bias neuron. Initialisation value within [ __offset , __offset + __range ] (offset value can be negative).

Parameters:
__network -

tanh

public static double tanh(double __value)
apply the hyperbolic tangent function. Useful for the eponyme activation function.

Parameters:
__value -
Returns: