NormalRandomDistribution< D, T > Class Template Reference

Random distribution for drawing samples from univariate or multivariate normal distributions. More...

#include <math/NormalRandomGenerator.h>

Inheritance diagram for NormalRandomDistribution< D, T >:

Public Types
typedef T	input_type
typedef Eigen::Matrix< T, D, 1 >	result_type
typedef detail::NormalRandomDistributionBase< T >	Base
typedef Eigen::Matrix< T, D, D >	VarianceType

Public Member Functions
	NormalRandomDistribution ()
	NormalRandomDistribution (const Eigen::Matrix< T, D, D > &sigma)
bool	setSigma (const Eigen::Matrix< T, D, D > &sigma)
	Sets the specified covariance matrix. More...
template<typename Engine >
Eigen::Matrix< T, D, 1 >	operator() (Engine &eng)
	Draws a sample from the normal distribution. More...

Detailed Description

template<int D, typename T = float>
class mira::NormalRandomDistribution< D, T >

Random distribution for drawing samples from univariate or multivariate normal distributions.

This class template models a zero-mean normal distribution of the specified dimension D (univariate if D==1 or multivariate if D>1).

It can be used with any of the boost random number generators:

boost::mt19937 engine;
NormalRandomDistribution<2,float> nrnd(Sigma);
boost::variate_generator<boost::mt19937&,    NormalRandomDistribution<2,float>> rnd(engine, nrnd);

// draw a sample:
Eigen::Vector2f sample = rnd();

The above instantiation is simplified by the NormalRandomGenerator.

Member Typedef Documentation

input_type

typedef T input_type

result_type

typedef Eigen::Matrix<T,D,1> result_type

Base

typedef detail::NormalRandomDistributionBase<T> Base

VarianceType

typedef Eigen::Matrix<T,D,D> VarianceType

Constructor & Destructor Documentation

NormalRandomDistribution() [1/2]

NormalRandomDistribution ( )

inline

NormalRandomDistribution() [2/2]

NormalRandomDistribution ( const Eigen::Matrix< T, D, D > &  sigma )

inline

Member Function Documentation

setSigma()

bool setSigma ( const Eigen::Matrix< T, D, D > &  sigma )

inline

Sets the specified covariance matrix.

The method assumes that the provided matrix is symmetric and uses the lower triangular part of the matrix only. The upper triangular part won't be read. If the matrix is not a positive definite matrix the member mL will not set and the method returns false, and otherwise true.

Internally a cholesky decomposition (LLT) of the covariance matrix is performed.

operator()()

Eigen::Matrix<T,D,1> operator() ( Engine &  eng )

inline

Draws a sample from the normal distribution.

---

The documentation for this class was generated from the following file:

• base/include/math/NormalRandomGenerator.h