# SuanShu, a Java numerical and statistical library

com.numericalmethod.suanshu.stats.distribution.multivariate.exponentialfamily

## Class MultivariateExponentialFamily

• java.lang.Object
• com.numericalmethod.suanshu.stats.distribution.multivariate.exponentialfamily.MultivariateExponentialFamily

• public class MultivariateExponentialFamily
extends Object
The exponential family is an important class of probability distributions sharing this particular form. $f_X(x|\boldsymbol \theta) = h(x) \exp\Big(\boldsymbol\eta({\boldsymbol \theta}) \cdot \mathbf{T}(x) - A({\boldsymbol \theta}) \Big)$
Wikipedia: Exponential family
• ### Constructor Summary

Constructors
Constructor and Description
MultivariateExponentialFamily(RealScalarFunction h, RealVectorFunction eta, RealVectorFunction T, RealScalarFunction A)
Construct a factory to construct probability distribution in the exponential family of this form.
• ### Method Summary

All Methods
Modifier and Type Method and Description
MultivariateProbabilityDistribution getDistribution(Vector theta)
Construct a probability distribution in the exponential family.
• ### Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
• ### Constructor Detail

• #### MultivariateExponentialFamily

public MultivariateExponentialFamily(RealScalarFunction h,
RealVectorFunction eta,
RealVectorFunction T,
RealScalarFunction A)
Construct a factory to construct probability distribution in the exponential family of this form. $f_X(x|\boldsymbol \theta) = h(x) \exp\Big(\boldsymbol\eta({\boldsymbol \theta}) \cdot \mathbf{T}(x) - A({\boldsymbol \theta}) \Big)$
Parameters:
h - the normalizing function
eta - the natural parameter
T - the sufficient statistic
A - the log-partition function
• ### Method Detail

• #### getDistribution

public MultivariateProbabilityDistribution getDistribution(Vector theta)
Construct a probability distribution in the exponential family.
Parameters:
theta - the parameter
Returns:
a fully specified probability distribution, $$f_X(x|\boldsymbol \theta)$$