mlpack::neighbor::DrusillaSelect< MatType > Class Template Reference

Public Member Functions
	DrusillaSelect (const MatType &referenceSet, const size_t l, const size_t m)
	Construct the DrusillaSelect object with the given reference set (this is the set that will be searched). More...
	DrusillaSelect (const size_t l, const size_t m)
	Construct the DrusillaSelect object with no given reference set. More...
void	Train (const MatType &referenceSet, const size_t l=0, const size_t m=0)
	Build the set of candidate points on the given reference set. More...
void	Search (const MatType &querySet, const size_t k, arma::Mat< size_t > &neighbors, arma::mat &distances)
	Search for the k furthest neighbors of the given query set. More...
template<typename Archive >
void	serialize (Archive &ar, const uint32_t)
	Serialize the model.
const MatType &	CandidateSet () const
	Access the candidate set.
MatType &	CandidateSet ()
	Modify the candidate set. Be careful!
const arma::Col< size_t > &	CandidateIndices () const
	Access the indices of points in the candidate set.
arma::Col< size_t > &	CandidateIndices ()
	Modify the indices of points in the candidate set. Be careful!

Constructor & Destructor Documentation

DrusillaSelect() [1/2]

template<typename MatType >

mlpack::neighbor::DrusillaSelect< MatType >::DrusillaSelect	(	const MatType &	referenceSet,
		const size_t	l,
		const size_t	m
	)

Construct the DrusillaSelect object with the given reference set (this is the set that will be searched).

The resulting set of candidate points that will be searched at query time will have size l*m.

Parameters

referenceSet	Set of reference data.
l	Number of projections.
m	Number of elements to store for each projection.

DrusillaSelect() [2/2]

template<typename MatType >

mlpack::neighbor::DrusillaSelect< MatType >::DrusillaSelect	(	const size_t	l,
		const size_t	m
	)

Construct the DrusillaSelect object with no given reference set.

Be sure to call Train() before calling Search()!

Parameters

l	Number of projections.
m	Number of elements to store for each projection.

Member Function Documentation

Search()

template<typename MatType >

void mlpack::neighbor::DrusillaSelect< MatType >::Search	(	const MatType &	querySet,
		const size_t	k,
		arma::Mat< size_t > &	neighbors,
		arma::mat &	distances
	)

Search for the k furthest neighbors of the given query set.

(The query set can contain just one point: that is okay.) The results will be stored in the given neighbors and distances matrices, in the same format as the NeighborSearch and LSHSearch classes. That is, each column in the neighbors and distances matrices will refer to a single query point, and the k'th row in that column will refer to the k'th candidate neighbor or distance for that query point.

Parameters

querySet	Set of query points to search.
k	Number of furthest neighbors to search for.
neighbors	Matrix to store resulting neighbors in.
distances	Matrix to store resulting distances in.

Train()

template<typename MatType >

void mlpack::neighbor::DrusillaSelect< MatType >::Train	(	const MatType &	referenceSet,
		const size_t	l = 0,
		const size_t	m = 0
	)

Build the set of candidate points on the given reference set.

If l and m are left unspecified, then the values set in the constructor will be used instead.

Parameters

referenceSet	Set to extract candidate points from.
l	Number of projections.
m	Number of elements to store for each projection.

---

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

• src/mlpack/methods/approx_kfn/drusilla_select.hpp
• src/mlpack/methods/approx_kfn/drusilla_select_impl.hpp