NearNeural

Class ThreePointRuleDifferentiation

java.lang.Object

NearNeural.ThreePointRuleDifferentiation

All Implemented Interfaces:

ThresholdingAlgorithm

public abstract class ThreePointRuleDifferentiation
extends java.lang.Object
implements ThresholdingAlgorithm

Implements a numerical method for differentiation. Extend this class if you cannot calculate a more accurate differentiation model for your thresholding function. This method takes copies of all edges in order to create small variations in value, then runs the value function two times - it is thus way over a guaranteed half speed of a well designed differentiation algorithm. Use this class when testing or in emergencies only, it is not fast and probably never will be.

Version:

0.2

Author:

Stephen Wattam

Constructor Summary

Constructors

Constructor and Description
ThreePointRuleDifferentiation(double p_h) Creates a new differentiation constant necessary to approximate differential.

Method Summary

Methods

Modifier and Type	Method and Description
double	differentialValue(Edge[] edges) Attempts to find the differential of any function by way of numerical methods.
abstract boolean	validateInput(double value) Validates input against the applicable range for this thresholding algorithm.
abstract double	value(Edge[] edges) Calculates Neuron value from all Edges and their source Neurons.

Methods inherited from class java.lang.Object

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

Constructor Detail

ThreePointRuleDifferentiation

public ThreePointRuleDifferentiation(double p_h)

Creates a new differentiation constant necessary to approximate differential. Generally the smaller h is the better the result.

Parameters:

p_h - The difference between samples.

Method Detail

value

public abstract double value(Edge[] edges)

Calculates Neuron value from all Edges and their source Neurons.

Specified by:

value in interface ThresholdingAlgorithm

Parameters:

edges - This neuron's edges

Returns:

This Neuron's value

validateInput

public abstract boolean validateInput(double value)

Validates input against the applicable range for this thresholding algorithm.

Specified by:

validateInput in interface ThresholdingAlgorithm

Parameters:

value - The value to check

Returns:

True if the value is applicable to run though without error

differentialValue

public double differentialValue(Edge[] edges)

Attempts to find the differential of any function by way of numerical methods.

Specified by:

differentialValue in interface ThresholdingAlgorithm

Parameters:

edges - All edges leading to this neurons

Returns:

The differential of the value of the arbitrary thresholding functioon