com.croftsoft.apps.cyborg
Class Activity

java.lang.Object
  com.croftsoft.apps.cyborg.Activity

public final class Activity
extends Object

Surrogate directional activity. Translated from SurrogateDirectionalActivityV2.xls (2004 October) by Lawrence J. Cauller, Ph.D., which was donated to the Public Domain by Dr. Cauller on 2004-12-01.

Since:

2004-10-19

Version:

2004-12-23

Author:

David Wallace Croft

Constructor Summary

Activity()

Method Summary

static double	computeNoiselessActivation(double preferredDirection, double actualDirection) Generates activation for a neuron without random noise.
static double	generateNoisyActivation(double preferredDirection, double actualDirection) Generates activation for a neuron with random noise.
static void	main(String[] args)
static double	normInv(double probability, double mean, double standardDeviation)
static double	predictDirection(double[] preferredDirections, double[] noisyActivations) Predicts the direction for this trial.
static void	printTemplateActivations(PrintStream printStream, double[] preferredDirections) Prints the template activations with one standard dev of noise.
static void	runTrials(PrintStream printStream, double[] preferredDirections) Runs the trials, gathers statistics, and prints the results.

Methods inherited from class java.lang.Object

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

Constructor Detail

Activity

public Activity()

Method Detail

main

public static void main(String[] args)
                 throws Exception

Throws:

Exception

printTemplateActivations

public static void printTemplateActivations(PrintStream printStream,
                                            double[] preferredDirections)
                                     throws Exception

Prints the template activations with one standard dev of noise.

Throws:

Exception

runTrials

public static void runTrials(PrintStream printStream,
                             double[] preferredDirections)
                      throws Exception

Runs the trials, gathers statistics, and prints the results.

Throws:

Exception

computeNoiselessActivation

public static double computeNoiselessActivation(double preferredDirection,
                                                double actualDirection)

Generates activation for a neuron without random noise.

generateNoisyActivation

public static double generateNoisyActivation(double preferredDirection,
                                             double actualDirection)
                                      throws Exception

Generates activation for a neuron with random noise.

Throws:

Exception

normInv

public static double normInv(double probability,
                             double mean,
                             double standardDeviation)
                      throws org.apache.commons.math.MathException

Throws:

org.apache.commons.math.MathException

predictDirection

public static double predictDirection(double[] preferredDirections,
                                      double[] noisyActivations)

Predicts the direction for this trial. Knowing only the preferred directions for each neuron and their current activations, predict the direction for this trial. This method is the heart of the algorithm.