bsoc.nn
Class Weighter

java.lang.Object
  |
  +--bsoc.nn.Weighter
public class Weighter
extends java.lang.Object
implements java.io.Serializable

The weighter is used to preprocess actiavtion values of preceeding neurons. Weighting is a multiplication of the incoming value with a value between min- and max factor. What factor is taken depends on the weight of the synapse. The weight of a synapse is a value out of the intervall [0 .. maxw].

f = maxf + ((maxf - minf) * (w / maxw))
f ... Factor
maxf ... maxFactor
minf ... minFactor
w ... weight
maxw ... maxWeight

The weighter is implemented as an array of values where result values can be calculated by accessing these values. This should speed up the calculation of neuronal nets.

See Also:
Serialized Form

Constructor Summary
Weighter()
           
Weighter(double minFactor, double maxFactor)
           
Weighter(float minFactor, float maxFactor)
           
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Constructor Detail

Weighter

public Weighter()

Weighter

public Weighter(float minFactor,
                float maxFactor)

Weighter

public Weighter(double minFactor,
                double maxFactor)