|
|||||||||
PREV CLASS NEXT CLASS | FRAMES NO FRAMES | ||||||||
SUMMARY: INNER | FIELD | CONSTR | METHOD | DETAIL: FIELD | CONSTR | METHOD |
java.lang.Object | +--saaps.ann.mlff.Network
This class implements the multi layer feed forward neural network. The network is made up of one or several layers, which in turn contain neurons. Each neuron is defined by its weights, bias, and transfer function.
The network can be stored to the file system and read from the file system.
To run the network one must first set the weights, biases, and transfer functions,
either explicitly or from reading a stored network. Then one specifies the input
as an saaps.ann.Data
object. After it has been run the output
can be received as an saaps.ann.Data
object.
Example:
This example assumes that we have an 2 element array ofLayer
objects calledlayers
. The should also be aData
object called inData.Network net = new Network(layers);
net.setInput(inData);
Data outData = net.getOutput();
Constructor Summary | |
Network()
Creates a new empty Network |
|
Network(Layer[] l)
Create a network from an array of layers. |
|
Network(java.lang.String name)
Create a network from a stored file. |
Method Summary | |
Data |
getOutput()
Get the output. |
void |
run()
Run the network. |
void |
setInput(Data inp)
Present the input data to the network. |
Methods inherited from class java.lang.Object |
clone,
equals,
finalize,
getClass,
hashCode,
notify,
notifyAll,
toString,
wait,
wait,
wait |
Constructor Detail |
public Network()
public Network(java.lang.String name)
name
- The name of the file.public Network(Layer[] l)
l
- The layersMethod Detail |
public void setInput(Data inp)
inp
- The input data.public void run()
public Data getOutput()
If the network has not been run after new input data has been presented this method will also run the network.
|
|||||||||
PREV CLASS NEXT CLASS | FRAMES NO FRAMES | ||||||||
SUMMARY: INNER | FIELD | CONSTR | METHOD | DETAIL: FIELD | CONSTR | METHOD |