com.datumbox.framework.core.machinelearning.ensemblelearning

Class FixedCombinationRules

java.lang.Object

com.datumbox.framework.core.machinelearning.ensemblelearning.FixedCombinationRules

public class FixedCombinationRules
extends java.lang.Object

Implementation of various Fixed combination rules which can be used in order to combine the responses of multiple classifiers. References: http://www2.cs.uh.edu/~ceick/ML/Topic12.ppt

Constructor Summary

Constructors

Constructor and Description
FixedCombinationRules()

Method Summary

Modifier and Type	Method and Description
static AssociativeArray	average(DataTable2D classifierClassProbabilityMatrix) Combines the responses of the classifiers by using estimating the average of the probabilities of their responses.
static AssociativeArray	majorityVote(DataTable2D classifierClassProbabilityMatrix) Combines the responses of the classifiers by summing the votes of each winner class.
static AssociativeArray	maximum(DataTable2D classifierClassProbabilityMatrix) Combines the responses of the classifiers by using selecting the maximum probability of each class.
static AssociativeArray	median(DataTable2D classifierClassProbabilityMatrix) Combines the responses of the classifiers by using estimating the median of the probabilities of their responses.
static AssociativeArray	minimum(DataTable2D classifierClassProbabilityMatrix) Combines the responses of the classifiers by using selecting the minimum probability of each class.
static AssociativeArray	product(DataTable2D classifierClassProbabilityMatrix) Combines the responses of the classifiers by using estimating the product of the probabilities of their responses.
static AssociativeArray	sum(DataTable2D classifierClassProbabilityMatrix) Combines the responses of the classifiers by using estimating the sum of the probabilities of their responses.
static AssociativeArray	weightedAverage(DataTable2D classifierClassProbabilityMatrix, AssociativeArray classifierWeights) Combines the responses of the classifiers by estimating the weighted average of the probabilities of their responses.

Methods inherited from class java.lang.Object

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

Constructor Detail

FixedCombinationRules

public FixedCombinationRules()

Method Detail

sum

public static AssociativeArray sum(DataTable2D classifierClassProbabilityMatrix)

Combines the responses of the classifiers by using estimating the sum of the probabilities of their responses.

Parameters:

classifierClassProbabilityMatrix -

Returns:

average

public static AssociativeArray average(DataTable2D classifierClassProbabilityMatrix)

Combines the responses of the classifiers by using estimating the average of the probabilities of their responses.

Parameters:

classifierClassProbabilityMatrix -

Returns:

weightedAverage

public static AssociativeArray weightedAverage(DataTable2D classifierClassProbabilityMatrix,
                                               AssociativeArray classifierWeights)

Combines the responses of the classifiers by estimating the weighted average of the probabilities of their responses.

Parameters:

classifierClassProbabilityMatrix -

classifierWeights -

Returns:

median

public static AssociativeArray median(DataTable2D classifierClassProbabilityMatrix)

Combines the responses of the classifiers by using estimating the median of the probabilities of their responses.

Parameters:

classifierClassProbabilityMatrix -

Returns:

maximum

public static AssociativeArray maximum(DataTable2D classifierClassProbabilityMatrix)

Combines the responses of the classifiers by using selecting the maximum probability of each class.

Parameters:

classifierClassProbabilityMatrix -

Returns:

minimum

public static AssociativeArray minimum(DataTable2D classifierClassProbabilityMatrix)

Combines the responses of the classifiers by using selecting the minimum probability of each class.

Parameters:

classifierClassProbabilityMatrix -

Returns:

product

public static AssociativeArray product(DataTable2D classifierClassProbabilityMatrix)

Combines the responses of the classifiers by using estimating the product of the probabilities of their responses.

Parameters:

classifierClassProbabilityMatrix -

Returns:

majorityVote

public static AssociativeArray majorityVote(DataTable2D classifierClassProbabilityMatrix)

Combines the responses of the classifiers by summing the votes of each winner class.

Parameters:

classifierClassProbabilityMatrix -

Returns: