bsoc.nn
Class NetConnector

java.lang.Object
  |
  +--bsoc.nn.NetConnector
Direct Known Subclasses:
AlphaConnector
public class NetConnector
extends java.lang.Object
implements java.io.Serializable

The netconnector allowes to define the structure of neuronal nets and parameters for synapses and neurons.

See Also:
Serialized Form

Constructor Summary
NetConnector()
          Generates a default net.
NetConnector(int seed, IntVector nodesPerLayer, IntVector connProbMatrix, Weighter wgt, Transfer trans)
          Allows to construct a net-connector with parameters to build a net.
 
Method Summary
 void connectNet(bsoc.nn.ToBeConnectedInterface net)
           
 void initLayers(bsoc.nn.ToBeConnectedInterface net)
           
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Constructor Detail

NetConnector

public NetConnector()
Generates a default net. Four layers with four neurons each. Mainly for test purposes.

NetConnector

public NetConnector(int seed,
                    IntVector nodesPerLayer,
                    IntVector connProbMatrix,
                    Weighter wgt,
                    Transfer trans)
Allows to construct a net-connector with parameters to build a net.
Parameters:
nodesPerLayer - A list of integer values spezifying the number of layers and the number of neurons per layer. The first entry spezifies the input layer.
connProbMatrix - A list of lists of integer values spezifying the probabillity of a connection of the neurons of a layer to be connected to another layer. Values must be in percent.
wgt - The weighter used to preprocess input values to synapses.
trans - The transfer function used in neurons.
Method Detail

initLayers

public void initLayers(bsoc.nn.ToBeConnectedInterface net)

connectNet

public void connectNet(bsoc.nn.ToBeConnectedInterface net)