net.sourceforge.cilib.neuralnetwork.generic.topologybuilders
Class GenericTopologyBuilder
java.lang.Object
net.sourceforge.cilib.neuralnetwork.generic.topologybuilders.GenericTopologyBuilder
- All Implemented Interfaces:
- Initializable
- Direct Known Subclasses:
- FFNNgenericTopologyBuilder
public abstract class GenericTopologyBuilder
- extends Object
- implements Initializable
- Author:
- stefanv
This class froms the base for building any NN topology. It provides a few basic tools such as contructing a
layered topology, either connected or not connected. This provides a base for createLayerList(). createLayerList()
and createNeuronPipePool() are implemented in sub classes and together are responsible for building a valid
topology (Builder Design Pattern).
Methods inherited from class java.lang.Object |
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait |
activationFunction
protected NeuronFunction activationFunction
outputActivationFunction
protected NeuronFunction outputActivationFunction
GenericTopologyBuilder
public GenericTopologyBuilder()
createLayerList
public abstract ArrayList<ArrayList<NeuronConfig>> createLayerList()
- Assemble the Neural Network
- Returns:
- the Layer List of NeuronConfig that represents the Neural Network
getActivationFunction
public NeuronFunction getActivationFunction()
- Get the type of NeuronFunction that is used to construct the Neural Network
- Returns:
- The NeuronFunction
setActivationFunction
public void setActivationFunction(NeuronFunction activationFunction)
- Set the type of NeuronFunction to use for the hidden layer(s) of the Neural Network
- Parameters:
activationFunction
- The activation function to use when constructing the Neural Network
getOutputActivationFunction
public NeuronFunction getOutputActivationFunction()
- Get the type of NeuronFunction that is used to construct the output layer Neural Network
- Returns:
- The NeuronFunction
setOutputActivationFunction
public void setOutputActivationFunction(NeuronFunction outputActivationFunction)
- The type of NeuronFunction to use for output neurons, this may differ from the type used in the hidden layer(s)
- Parameters:
outputActivationFunction
- The activation funciton to use for output Neurons when construction the Neural Network
Copyright © 2009 CIRG. All Rights Reserved.