jsat.classifiers.neuralnetwork

Enum RBFNet.Phase2Learner

java.lang.Object

java.lang.Enum<RBFNet.Phase2Learner>

jsat.classifiers.neuralnetwork.RBFNet.Phase2Learner

All Implemented Interfaces:

Serializable, Comparable<RBFNet.Phase2Learner>

Enclosing class:

RBFNet

public static enum RBFNet.Phase2Learner
extends Enum<RBFNet.Phase2Learner>

The second phase of learning a RBF Neural Network is to determine how the neurons are activated to produce the output of the hidden layer. This enum control which method is used.

Enum Constant Summary

Enum Constants

Enum Constant and Description
CENTROID_DISTANCE This method sets the bandwidth for each neuron based on the distances to the neuron from each data point that is closest to said neuron.
CLOSEST_OPPOSITE_CENTROID This bandwidth estimator only works for classification problems.
NEAREST_OTHER_CENTROID_AVERAGE This method sets the bandwidth for each neuron based on the average distance of the p nearest neurons.

Method Summary

Modifier and Type	Method and Description
protected abstract double[]	estimateBandwidths(double alpha, int p, DataSet data, List<Vec> centroids, List<Double> centroidDistCache, DistanceMetric dm, ExecutorService threadpool)
static RBFNet.Phase2Learner	valueOf(String name) Returns the enum constant of this type with the specified name.
static RBFNet.Phase2Learner[]	values() Returns an array containing the constants of this enum type, in the order they are declared.

Methods inherited from class java.lang.Enum

clone, compareTo, equals, finalize, getDeclaringClass, hashCode, name, ordinal, toString, valueOf

Methods inherited from class java.lang.Object

getClass, notify, notifyAll, wait, wait, wait

Enum Constant Detail

CENTROID_DISTANCE

public static final RBFNet.Phase2Learner CENTROID_DISTANCE

This method sets the bandwidth for each neuron based on the distances to the neuron from each data point that is closest to said neuron. If μ is the average distance to the neuron, and σ the standard deviation, then the bandwidth b of the j'th neuron is seto to b_j = μ_j + α σ_j

CLOSEST_OPPOSITE_CENTROID

public static final RBFNet.Phase2Learner CLOSEST_OPPOSITE_CENTROID

This bandwidth estimator only works for classification problems. Each neuron is assigned a class based on the majority class labels of the data points closes to said neuron. The bandwidth is then estimated as α times the distance from the neuron to the closest neuron with a different class label.

For this method α values between (0, 1) usually work best, 0.25 is a good starting value. The value of α can go past 1.

NEAREST_OTHER_CENTROID_AVERAGE

public static final RBFNet.Phase2Learner NEAREST_OTHER_CENTROID_AVERAGE

This method sets the bandwidth for each neuron based on the average distance of the p nearest neurons. The number of standard deviations to add to the activation is controlled by α

Method Detail

values

public static RBFNet.Phase2Learner[] values()

Returns an array containing the constants of this enum type, in the order they are declared. This method may be used to iterate over the constants as follows:

for (RBFNet.Phase2Learner c : RBFNet.Phase2Learner.values())
    System.out.println(c);

Returns:

an array containing the constants of this enum type, in the order they are declared

valueOf

public static RBFNet.Phase2Learner valueOf(String name)

Returns the enum constant of this type with the specified name. The string must match exactly an identifier used to declare an enum constant in this type. (Extraneous whitespace characters are not permitted.)

Parameters:

name - the name of the enum constant to be returned.

Returns:

the enum constant with the specified name

Throws:

IllegalArgumentException - if this enum type has no constant with the specified name

NullPointerException - if the argument is null

estimateBandwidths

protected abstract double[] estimateBandwidths(double alpha,
                                               int p,
                                               DataSet data,
                                               List<Vec> centroids,
                                               List<Double> centroidDistCache,
                                               DistanceMetric dm,
                                               ExecutorService threadpool)