Documentation of 'org.joone.helpers.structure.NeuralNetMatrix' Java class
NeuralNetMatrix
Skip navigation links
org.joone.helpers.structure

Class NeuralNetMatrix


  • public class NeuralNetMatrix
extends java.lang.Object
    Utility class that performs several useful 'decompositions' of a NeuralNet object, making very simple to handle its internal structure. This class can be useful to analyze or transfor a neural network, because it transforms the network to a 'flat' structure, where there isn't any object pointed more than once. It, also, by means of the getConnectionMatrix public method, returns a 2D representation of the internal Synapses of a neural network.

    • Constructor Detail

      • NeuralNetMatrix

        public NeuralNetMatrix()
        Creates a new instance of NeuralNetMatrix

      • NeuralNetMatrix

        public NeuralNetMatrix(NeuralNet net)
        Creates a new instance of NeuralNetMatrix

    • Method Detail

      • setNeuralNet

        public void setNeuralNet(NeuralNet net)
        Method used to set the neural network to dissect

      • getConnectionMatrix

        public Synapse[][] getConnectionMatrix()
        Converts the neural structure to a matrix containing all the synapses of the neural network. At each not-null [x][y] element of the 2D array, there is a pointer to the Synapse connecting the Layer[x] to the Layer[y]. The returned 2D array of Synapses could be used, for instance, to draw a graphical representation of the neural network.

      • isThereAnyPath

        public boolean isThereAnyPath(Layer fromLayer,
                              Layer toLayer)
        Searches a path between two layers. Useful in order to discover recurrent networks
        Returns:
        true if there is a path from fromLayer to toLayer

      • getOrderedConnectionMatrix

        public Synapse[][] getOrderedConnectionMatrix()
        Converts the neural structure to a matrix containing all the synapses of the neural network. The indexes indicates the layer in Layer[] getOrderedLayers(). At each not-null [x][y] element of the 2D array, there is a pointer to the Synapse connecting the (ordered)Layer[x] to the (ordered)Layer[y]. The returned 2D array of Synapses could be used, for instance, to draw a graphical representation of the neural network.
        Returns:
        a matrix where the indexes x,y point to the Layers returned by getOrderedLayers().

      • getBinaryOrderedConnectionMatrix

        public boolean[][] getBinaryOrderedConnectionMatrix()
        Converts the neural structure to a matrix containing all the synapses of the neural network. The index indicates the layer in Layer[] getOrderedLayers(). True means connection between the corresponding layers.
        Returns:
        a matrix where the indexes x,y point to the Layers returned by getOrderedLayers().

      • getOrderedLayers

        public Layer[] getOrderedLayers()
        This method calculates the order of the layers of a neural network, from the input to the output. This method fills also the translations array.
        Returns:
        An array containing the ordered Layers, from the input to the output (i.e. layers[0]=input layer, layers[n-1]=output layer.

      • getInputLayers

        public java.util.ArrayList getInputLayers(Synapse[][] connMatrix)
        Searches for all the input layers of the network. An input layer is represented by each column in connectionMatrix that doesn't contain any Synapse
        Returns:
        an ArrayList containing Integers that point to the indexes of the input layers

      • cloneElement

        public java.io.Serializable cloneElement(java.io.Serializable element)
        Clones a neural element.
        Parameters:
        element - The element to clone
        Returns:
        the clone of the element passed as parameter

      • getLayers

        public java.util.ArrayList getLayers()
        Getter for property layers.
        Returns:
        Value of property layers.

      • setLayers

        public void setLayers(java.util.ArrayList layers)
        Setter for property layers.
        Parameters:
        layers - New value of property layers.

      • getConnectionSet

        public java.util.ArrayList getConnectionSet()
        Getter for property connectionSet.
        Returns:
        Value of property connectionSet.

      • setConnectionSet

        public void setConnectionSet(java.util.ArrayList connectionSet)
        Setter for property connectionSet.
        Parameters:
        connectionSet - New value of property connectionSet.

      • getMonitor

        public Monitor getMonitor()
        Getter for property monitor.
        Returns:
        Value of property monitor.

      • setMonitor

        public void setMonitor(Monitor monitor)
        Setter for property monitor.
        Parameters:
        monitor - New value of property monitor.

      • getInputLayer

        public Layer getInputLayer()

      • setInputLayer

        public void setInputLayer(Layer inputLayer)

      • getOutputLayer

        public Layer getOutputLayer()

      • setOutputLayer

        public void setOutputLayer(Layer outputLayer)

      • getNumLayers

        public int getNumLayers()

      • getInputLayerInd

        public int getInputLayerInd()
        Calculates the index of the input layer
        Returns:
        The index, within NeuralNet.layers[], of the input layer

      • getOutputLayerInd

        public int getOutputLayerInd()

      • getLayerInd

        public int getLayerInd(Layer layer)
        Calculates the index of a layer within the layers array. This method uses the NeuralNet's
        Returns:
        the Layer's index starting from 0. Returns -1 if not found
Skip navigation links

DataMelt 3.0 © DataMelt by jWork.ORG

You see the box below because you did not login.