Infer Poissonian mixed model. More...

#include <yat/statistics/PoissonMixture.h>

Public Member Functions
void	add (unsigned long int k, unsigned long int n=1)

void	clear (void)
	Remove all data points.

void	fit (size_t n)

void	fit (const yat::utility::VectorBase &m, const yat::utility::VectorBase &tau)

double	logL (void)

double	mean (size_t i) const

double	tau (size_t i) const

Detailed Description

Infer Poissonian mixed model.

Data are described as a mixed model of Poissonians $P(X=k) \sum \tau_i po_i(k)$ where $po_i(k) = \frac{m^k e^{-m}}{k!}$

Since: New in yat 0.20

Member Function Documentation

◆ add()

void theplu::yat::statistics::PoissonMixture::add	(	unsigned long int	k,
		unsigned long int	n = `1`
	)

Add n data point(s)

◆ fit() [1/2]

void theplu::yat::statistics::PoissonMixture::fit ( size_t n )

Create n sub-models and tune their parameters such that the logL is maximized.

◆ fit() [2/2]

void theplu::yat::statistics::PoissonMixture::fit	(	const yat::utility::VectorBase &	m,
		const yat::utility::VectorBase &	tau
	)

Initialise the model parameters as m and tau and fit the data by maximizing logL.

◆ logL()

double theplu::yat::statistics::PoissonMixture::logL ( void )

Log likelihood is calculated as $\ln L = \sum_s \ln \left( \sum \tau_i po_i(k_s) \right) = \sum_s \ln \left( \sum \tau_i \frac{m_i^{k_s} e^{-m_i}}{k_s!} \right) =$

◆ mean()

double theplu::yat::statistics::PoissonMixture::mean ( size_t i ) const

Returns: mean of model i

◆ tau()

double theplu::yat::statistics::PoissonMixture::tau ( size_t i ) const

Returns: tau corresponding to model i

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

yat/statistics/PoissonMixture.h

Public Member Functions