Class ContextNeuron

java.lang.Object
  Neuron
      ContextNeuron

public class ContextNeuron

extends Neuron

The ContextNeuron class creates Context Neuron which are used to provide recurrency in the RNN network. They store the previous Hidden Neuron activated values and a variable amount of their own previous values, which is fed back into the network

Author:

Raymond McBride

See Also:

RNN

Constructor Summary

ContextNeuron(double memoryDepth)
This constructor for the ContextNeuron specifies the memory depth and sets the input value to 0.5

Method Summary

void	calculateDelayedOutput() Calculates the summation of the delayed output values
void	calculateOutput() Calculates the output value
protected double	getInputValue() Gets the input value
protected double	getOutputValue() Gets the output value
void	input(double input) Set the new input value and adds the old value to the memory
void	setBias(double input) Sets the biased input
protected void	setInputValue(double inputValue) Sets the input value
protected void	setOutputValue(double outputValue) Sets the output value
protected double	sigmoidActivation(double summation) The Sigmoid Activation Function

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

ContextNeuron

public ContextNeuron(double memoryDepth)

This constructor for the ContextNeuron specifies the memory depth and sets the input value to 0.5

Parameters:

memoryDepth - The memory depth

Method Detail

input

public void input(double input)

Set the new input value and adds the old value to the memory

Parameters:

input - The new input value

calculateOutput

public void calculateOutput()

Calculates the output value

setBias

public void setBias(double input)

Sets the biased input

Parameters:

input - The new input value

calculateDelayedOutput

public void calculateDelayedOutput()

Calculates the summation of the delayed output values

sigmoidActivation

protected double sigmoidActivation(double summation)

The Sigmoid Activation Function

Returns:

the activated value

getInputValue

protected double getInputValue()

Gets the input value

Returns:

the input value

setInputValue

protected void setInputValue(double inputValue)

Sets the input value

Parameters:

inputValue - the input value

getOutputValue

protected double getOutputValue()

Gets the output value

Returns:

the output value

setOutputValue

protected void setOutputValue(double outputValue)

Sets the output value

Parameters:

outputValue - the output value