JMSLTM Numerical Library 4.0

com.imsl.datamining.neural
Class Link

java.lang.Object
  extended bycom.imsl.datamining.neural.Link
All Implemented Interfaces:
Serializable

public class Link
extends Object
implements Serializable

A link in a neural network.

Link objects are not created directly. Instead, they are created by factory methods in FeedForwardNetwork.

The most useful method is FeedForwardNetwork.linkAll(com.imsl.datamining.neural.Layer, com.imsl.datamining.neural.Layer) which creates Link objects connecting every Node in each Layer to every Node in the next Layer .

The method FeedForwardNetwork.link(Node,Node) creates a Link from a Node to any Node in a later Layer.

The method FeedForwardNetwork.findLink(Node,Node) returns the Link connecting two Nodes in the Network.

The method FeedForwardNetwork.remove(Link) removes a Link from the Network.

Each Link object contains a weight. Weights are used in computing Perceptron values.

See Also:
FeedForwardNetwork, Serialized Form

Method Summary
 Node getFrom()
          Returns the origination Node for this Link.
 Node getTo()
          Returns the destination Node for this Link.
 double getWeight()
          Returns the weight for this Link.
 void setWeight(double weight)
          Sets the weight for this Link.
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Method Detail

getFrom

public Node getFrom()
Returns the origination Node for this Link.

Returns:
A Node which is the origination Node for this Link.

getTo

public Node getTo()
Returns the destination Node for this Link.

Returns:
A Node which is the destination Node for this Link.

getWeight

public double getWeight()
Returns the weight for this Link.

Returns:
A double which contains the weight attributed to this Node.

setWeight

public void setWeight(double weight)
Sets the weight for this Link.

Parameters:
weight - A double which specifies the weight to attribute to this Link.

JMSLTM Numerical Library 4.0

Copyright 1970-2006 Visual Numerics, Inc.
Built June 1 2006.