KJB
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Macros Groups Pages
Public Member Functions | List of all members
kjb::mcmcda::Likelihood< Track > Class Template Reference

Computes the GP-based likelihood of an association. More...

#include <mcmcda_likelihood.h>

Public Member Functions

 Likelihood (double scale, double signal_sigma, double noise_sigma, const typename Data< Element >::Convert &to_vector)
 Constructor. More...
 
double operator() (const Association< Track > &w) const
 Applies this functor to the given association. More...
 
double at_noise_point (const Element &pt) const
 Returns the likelihood of the noise track. More...
 
double at_noise (const Available_data &false_alarms) const
 Returns the likelihood of the noise track. More...
 
double at_track (const Track &track) const
 Computes the GP log-likelihood of a track. More...
 
double get_noise_sigma () const
 Return the noise sigma of this model. More...
 
const Gp & get_gp () const
 Return the smoothness scale of this model. More...
 
const Data< Element >::Convert & get_convert () const
 Return the convert function. More...
 
void fix_inputs (const Gp_inputs &ins, size_t dim) const
 Fixes the inputs for faster likelihood computation. More...
 
void unfix_inputs () const
 Fixes the inputs for faster likelihood computation. More...
 
void set_limits (int low, int up) const
 Set limits on evaluation. More...
 
void reset_limits () const
 Unser limits on evaluation. More...
 

Detailed Description

template<class Track>
class kjb::mcmcda::Likelihood< Track >

Computes the GP-based likelihood of an association.

This functor computes the log-likelihood of an association, based on a Gaussian process model of motion. Essentially, it computes the marginal log-likelihood of the data of every track and adds them together. It also computes the likelihood of the noise (the unused data).

Constructor & Destructor Documentation

template<class Track >
kjb::mcmcda::Likelihood< Track >::Likelihood ( double  scale,
double  signal_sigma,
double  noise_sigma,
const typename Data< Element >::Convert &  to_vector 
)
inline

Constructor.

Parameters
noise_sigmaThe variance of the noise process
scaleThe scale parameter of the GP.

Member Function Documentation

template<class Track >
double kjb::mcmcda::Likelihood< Track >::at_noise ( const Available_data &  false_alarms) const

Returns the likelihood of the noise track.

template<class Track >
double kjb::mcmcda::Likelihood< Track >::at_noise_point ( const Element &  pt) const
inline

Returns the likelihood of the noise track.

template<class Track >
double kjb::mcmcda::Likelihood< Track >::at_track ( const Track &  track) const

Computes the GP log-likelihood of a track.

template<class Track >
void kjb::mcmcda::Likelihood< Track >::fix_inputs ( const Gp_inputs &  ins,
size_t  dim 
) const
inline

Fixes the inputs for faster likelihood computation.

template<class Track >
const Data<Element>::Convert& kjb::mcmcda::Likelihood< Track >::get_convert ( ) const
inline

Return the convert function.

template<class Track >
const Gp& kjb::mcmcda::Likelihood< Track >::get_gp ( ) const
inline

Return the smoothness scale of this model.

template<class Track >
double kjb::mcmcda::Likelihood< Track >::get_noise_sigma ( ) const
inline

Return the noise sigma of this model.

template<class Track >
double kjb::mcmcda::Likelihood< Track >::operator() ( const Association< Track > &  w) const

Applies this functor to the given association.

Returns
The log-likelihood of the association
template<class Track >
void kjb::mcmcda::Likelihood< Track >::reset_limits ( ) const
inline

Unser limits on evaluation.

template<class Track >
void kjb::mcmcda::Likelihood< Track >::set_limits ( int  low,
int  up 
) const
inline

Set limits on evaluation.

template<class Track >
void kjb::mcmcda::Likelihood< Track >::unfix_inputs ( ) const
inline

Fixes the inputs for faster likelihood computation.


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