dirichlet_multinomial_rng()

template<class RNG >

std::vector< int > stan::math::dirichlet_multinomial_rng ( const Eigen::Matrix< double, Eigen::Dynamic, 1 > &  alpha, int  N, RNG &  rng  )

inline

Return a draw from a Dirichlet-Multinomial distribution with specified parameters \(\alpha\) and \(N\) and pseudo-random number generator rng.

The Dirichlet-Multinomial distribution is a continuous mixture of Multinomial distirbutions, where the mixing distribution is the Dirichlet distribution. This fact is used for generating DirMult random draws. First, we sample a probability vector \(\theta \sim \mbox{Dirichlet}(\alpha)\). Then, we sample a \(n \sim \mbox{Multinomial}(\theta, N)\), and return \(n\).

Template Parameters

RNG	type of pseudo-random number generator

Parameters

alpha	Prior sample sizes (or intensity vector).
N	Number of trials.
rng	Pseudo-random number generator.

Returns

A non-negative integer vector n with sum(n) = N.

Exceptions

std::domain_error	if any element of alpha is less than or equal to 0, or infinite.
std::domain_error	is N is less than 0.

Definition at line 37 of file dirichlet_multinomial_rng.hpp.