saaps.network
Class Network

java.lang.Object
  |
  +--saaps.network.Network

public class Network
extends java.lang.Object

The Network class consist of one or several Layer objects.

Version:
2000-01-19
Author:
Peter Wintoft

Field Summary
private  Layer[] layer
           
private  Jama.Matrix output
           
 
Constructor Summary
Network(Jama.Matrix w1, Jama.Matrix b1)
          Construct a one-layer network using the default activation function (TANSIG).
Network(Jama.Matrix w1, Jama.Matrix b1, int f1)
          Construct a one-layer network.
Network(Jama.Matrix w1, Jama.Matrix b1, int f1, Jama.Matrix w2, Jama.Matrix b2, int f2)
          Construct a two-layer network.
Network(Jama.Matrix w1, Jama.Matrix b1, Jama.Matrix w2, Jama.Matrix b2)
          Construct a two-layer network using the default activation function (TANSIG).
Network(NetworkParam np)
          Construct a network from a NetworkParam object.
 
Method Summary
 Jama.Matrix getOutput()
          Get the output from the network.
 Jama.Matrix sim(Jama.Matrix input)
          Run the network from an input matrix.
 
Methods inherited from class java.lang.Object
, clone, equals, finalize, getClass, hashCode, notify, notifyAll, registerNatives, toString, wait, wait, wait
 

Field Detail

layer

private Layer[] layer

output

private Jama.Matrix output
Constructor Detail

Network

public Network(NetworkParam np)
Construct a network from a NetworkParam object.
Parameters:
np - the NetworkParam object

Network

public Network(Jama.Matrix w1,
               Jama.Matrix b1,
               int f1)
Construct a one-layer network.

May be removed in the future.

Parameters:
w1 - weights
b1 - bias
f1 - activation function

Network

public Network(Jama.Matrix w1,
               Jama.Matrix b1)
Construct a one-layer network using the default activation function (TANSIG).

May be removed in the future.

Parameters:
w1 - weights
b1 - bias

Network

public Network(Jama.Matrix w1,
               Jama.Matrix b1,
               int f1,
               Jama.Matrix w2,
               Jama.Matrix b2,
               int f2)
Construct a two-layer network.

May be removed in the future.

Parameters:
w1 - layer one weights
b1 - layer one bias
f1 - layer one activation function
w2 - layer two weights
b2 - layer two bias
f2 - layer two activation function

Network

public Network(Jama.Matrix w1,
               Jama.Matrix b1,
               Jama.Matrix w2,
               Jama.Matrix b2)
Construct a two-layer network using the default activation function (TANSIG).

May be removed in the future.

Parameters:
w1 - layer one weights
b1 - layer one bias
w2 - layer two weights
b2 - layer two bias
Method Detail

sim

public Jama.Matrix sim(Jama.Matrix input)
Run the network from an input matrix. The output will be a Matrix object with the number of rows equal to the number of outputs and the number of columns equal to the number of columns in the input matrix.
Parameters:
input - the input matrix

getOutput

public Jama.Matrix getOutput()
Get the output from the network.