cc.glsn.v15.neuralnet

Class Neuron

java.lang.Object

cc.glsn.v15.neuralnet.Neuron

public class Neuron
extends Object

Behold the noble neuron. It takes the summation of the product of the input values times their weights and runs that through the activation function (g). Put another way: output = g( sum(W_i * o_i) ) sum() sum over all inputs W_i - weight of input i o_i - value from input i There is no reason for a user of this package to muck directly with neurons. They should be used via the Brain class. However, I've left the Neuron class unprotected in case someone wants to use them directly.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
Neuron(NetFunction g, Random r)

Method Summary

Methods

Modifier and Type	Method and Description
void	addToBackProp(double output_error) Thanks to Paras Chopra (www.paraschopra.com) for providing sample code that I was able to use to get a handle on the back propogation
void	doAccumulatedBackProp(double alpha) Thanks to Paras Chopra (www.paraschopra.com) for providing sample code that I was able to use to get a handle on the back propogation
double	getSum(long state_serial) Get the output of the sumation of all the inputs*weights before the translation function is run.
double	getValue(long state_serial) Get the output value
void	setSources(cc.glsn.v15.neuralnet.NetworkSource[] srcs)
String	toString()

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, wait, wait, wait

Constructor Detail

Neuron

public Neuron(NetFunction g,
      Random r)

Method Detail

getValue

public double getValue(long state_serial)

Get the output value

Parameters:

state_serial -

Returns:

the output for this neuron

getSum

public double getSum(long state_serial)

Get the output of the sumation of all the inputs*weights before the translation function is run.

Parameters:

state_serial -

Returns:

the sum

setSources

public void setSources(cc.glsn.v15.neuralnet.NetworkSource[] srcs)

addToBackProp

public void addToBackProp(double output_error)

Thanks to Paras Chopra (www.paraschopra.com) for providing sample code that I was able to use to get a handle on the back propogation

doAccumulatedBackProp

public void doAccumulatedBackProp(double alpha)

Thanks to Paras Chopra (www.paraschopra.com) for providing sample code that I was able to use to get a handle on the back propogation

toString

public String toString()

Overrides:

toString in class Object