Package org.tribuo.math.distributions

Class MultivariateNormalDistribution

java.lang.Object

org.tribuo.math.distributions.MultivariateNormalDistribution

public final class MultivariateNormalDistribution
extends Object

A class for sampling from multivariate normal distributions.

Constructor Summary

Constructors

Constructor	Description
MultivariateNormalDistribution(double[] means, double[][] covariance, long seed)	Constructs a multivariate normal distribution that can be sampled from.
MultivariateNormalDistribution(double[] means, double[][] covariance, long seed, boolean eigenDecomposition)	Constructs a multivariate normal distribution that can be sampled from.
MultivariateNormalDistribution(DenseVector means, DenseMatrix covariance, long seed)	Constructs a multivariate normal distribution that can be sampled from.
MultivariateNormalDistribution(DenseVector means, DenseMatrix covariance, long seed, boolean eigenDecomposition)	Constructs a multivariate normal distribution that can be sampled from.

Method Summary

Modifier and Type	Method	Description
double[]	sampleArray()	Sample a vector from this multivariate normal distribution.
DenseVector	sampleVector()	Sample a vector from this multivariate normal distribution.
String	toString()

Methods inherited from class java.lang.Object

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

Constructor Details

MultivariateNormalDistribution

public MultivariateNormalDistribution(double[] means,
 double[][] covariance,
 long seed)

Constructs a multivariate normal distribution that can be sampled from.

Throws IllegalArgumentException if the covariance matrix is not positive definite.

Uses a DenseMatrix.CholeskyFactorization to compute the sampling covariance matrix.

Parameters:

means - The mean vector.

covariance - The covariance matrix.

seed - The RNG seed.

MultivariateNormalDistribution

public MultivariateNormalDistribution(double[] means,
 double[][] covariance,
 long seed,
 boolean eigenDecomposition)

Constructs a multivariate normal distribution that can be sampled from.

Throws IllegalArgumentException if the covariance matrix is not positive definite.

Parameters:

means - The mean vector.

covariance - The covariance matrix.

seed - The RNG seed.

eigenDecomposition - If true use an eigen decomposition to compute the sampling covariance matrix rather than a cholesky factorization.

MultivariateNormalDistribution

public MultivariateNormalDistribution(DenseVector means,
 DenseMatrix covariance,
 long seed)

Constructs a multivariate normal distribution that can be sampled from.

Throws IllegalArgumentException if the covariance matrix is not positive definite.

Uses a DenseMatrix.CholeskyFactorization to compute the sampling covariance matrix.

Parameters:

means - The mean vector.

covariance - The covariance matrix.

seed - The RNG seed.

MultivariateNormalDistribution

public MultivariateNormalDistribution(DenseVector means,
 DenseMatrix covariance,
 long seed,
 boolean eigenDecomposition)

Constructs a multivariate normal distribution that can be sampled from.

Throws IllegalArgumentException if the covariance matrix is not positive definite.

Parameters:

means - The mean vector.

covariance - The covariance matrix.

seed - The RNG seed.

eigenDecomposition - If true use an eigen decomposition to compute the sampling covariance matrix rather than a cholesky factorization.

Method Details

sampleVector

public DenseVector sampleVector()

Sample a vector from this multivariate normal distribution.

Returns:

A sample from this distribution.

sampleArray

public double[] sampleArray()

Sample a vector from this multivariate normal distribution.

Returns:

A sample from this distribution.

toString

public String toString()

Overrides:

toString in class Object