org.apache.ignite.ml.math.stat

Class DistributionMixture<C extends Distribution>

java.lang.Object

org.apache.ignite.ml.math.stat.DistributionMixture<C>

Type Parameters:

C - distributions mixture component class.

All Implemented Interfaces:

Serializable, Distribution

Direct Known Subclasses:

GmmModel

public abstract class DistributionMixture<C extends Distribution>
extends Object
implements Distribution

Mixture of distributions class where each component has own probability and probability of input vector can be computed as a sum of likelihoods of each component.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
DistributionMixture(Vector componentProbs, List<C> distributions) Creates an instance of DistributionMixture.

Method Summary

Modifier and Type	Method and Description
Vector	componentsProbs()
int	countOfComponents()
int	dimension()
List<C>	distributions()
Vector	likelihood(Vector x)
double	prob(Vector x)

Methods inherited from class java.lang.Object

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

Constructor Detail

DistributionMixture

public DistributionMixture(Vector componentProbs,
                           List<C> distributions)

Creates an instance of DistributionMixture.

Parameters:

componentProbs - Component probabilities.

distributions - Distributions.

Method Detail

prob

public double prob(Vector x)

Specified by:

prob in interface Distribution

Parameters:

x - Vector.

Returns:

Probability of vector.

likelihood

public Vector likelihood(Vector x)

Parameters:

x - Vector.

Returns:

Vector consists of likelihoods of each mixture components.

countOfComponents

public int countOfComponents()

Returns:

An amount of components.

componentsProbs

public Vector componentsProbs()

Returns:

Component probabilities.

distributions

public List<C> distributions()

Returns:

List of components.

dimension

public int dimension()

Specified by:

dimension in interface Distribution

Returns:

Dimension of vector space.