mlpack  2.2.4
mlpack::neighbor Namespace Reference

Neighbor-search routines. More...

Classes

class  BiSearchVisitor
 BiSearchVisitor executes a bichromatic neighbor search on the given NSType. More...

 
class  DeleteVisitor
 DeleteVisitor deletes the given NSType instance. More...

 
class  DrusillaSelect
 
class  EpsilonVisitor
 EpsilonVisitor exposes the Epsilon method of the given NSType. More...

 
class  FurthestNeighborSort
 This class implements the necessary methods for the SortPolicy template parameter of the NeighborSearch class. More...

 
class  LSHSearch
 The LSHSearch class; this class builds a hash on the reference set and uses this hash to compute the distance-approximate nearest-neighbors of the given queries. More...

 
class  MonoSearchVisitor
 MonoSearchVisitor executes a monochromatic neighbor search on the given NSType. More...

 
class  NearestNeighborSort
 This class implements the necessary methods for the SortPolicy template parameter of the NeighborSearch class. More...

 
class  NeighborSearch
 The NeighborSearch class is a template class for performing distance-based neighbor searches. More...

 
class  NeighborSearchRules
 The NeighborSearchRules class is a template helper class used by NeighborSearch class when performing distance-based neighbor searches. More...

 
class  NeighborSearchStat
 Extra data for each node in the tree. More...

 
class  NSModel
 The NSModel class provides an easy way to serialize a model, abstracts away the different types of trees, and also reflects the NeighborSearch API. More...

 
struct  NSModelName
 
struct  NSModelName< FurthestNeighborSort >
 
struct  NSModelName< NearestNeighborSort >
 
class  QDAFN
 
class  RAModel
 The RAModel class provides an abstraction for the RASearch class, abstracting away the TreeType parameter and allowing it to be specified at runtime in this class. More...

 
class  RAQueryStat
 Extra data for each node in the tree. More...

 
class  RASearch
 The RASearch class: This class provides a generic manner to perform rank-approximate search via random-sampling. More...

 
class  RASearchRules
 The RASearchRules class is a template helper class used by RASearch class when performing rank-approximate search via random-sampling. More...

 
class  RAUtil
 
class  ReferenceSetVisitor
 ReferenceSetVisitor exposes the referenceSet of the given NSType. More...

 
class  SearchModeVisitor
 SearchModeVisitor exposes the SearchMode() method of the given NSType. More...

 
class  SetSearchModeVisitor
 SetSearchModeVisitor modifies the SearchMode method of the given NSType. More...

 
class  TrainVisitor
 TrainVisitor sets the reference set to a new reference set on the given NSType. More...

 

Typedefs

typedef NeighborSearch< FurthestNeighborSort, metric::EuclideanDistanceAllkFN
 
typedef NeighborSearch< NearestNeighborSort, metric::EuclideanDistanceAllkNN
 
typedef RASearch AllkRAFN
 
typedef RASearch AllkRANN
 
template
<
template
<
typename
TreeMetricType
,
typename
TreeStatType
,
typename
TreeMatType
>
class
TreeType
=
tree::SPTree
>
using DefeatistKNN = NeighborSearch< NearestNeighborSort, metric::EuclideanDistance, arma::mat, TreeType, TreeType< metric::EuclideanDistance, NeighborSearchStat< NearestNeighborSort >, arma::mat >::template DefeatistDualTreeTraverser, TreeType< metric::EuclideanDistance, NeighborSearchStat< NearestNeighborSort >, arma::mat >::template DefeatistSingleTreeTraverser >
 The DefeatistKNN class is the k-nearest-neighbors method considering defeatist search. More...

 
typedef NeighborSearch< FurthestNeighborSort, metric::EuclideanDistanceKFN
 The KFN class is the k-furthest-neighbors method. More...

 
typedef NeighborSearch< NearestNeighborSort, metric::EuclideanDistanceKNN
 The KNN class is the k-nearest-neighbors method. More...

 
typedef RASearch< FurthestNeighborSortKRAFN
 The KRAFN class is the k-rank-approximate-farthest-neighbors method. More...

 
typedef RASearch KRANN
 The KRANN class is the k-rank-approximate-nearest-neighbors method. More...

 
template
<
typename
SortPolicy
,
template
<
typename
TreeMetricType
,
typename
TreeStatType
,
typename
TreeMatType
>
class
TreeType
>
using NSType = NeighborSearch< SortPolicy, metric::EuclideanDistance, arma::mat, TreeType, TreeType< metric::EuclideanDistance, NeighborSearchStat< SortPolicy >, arma::mat >::template DualTreeTraverser >
 Alias template for euclidean neighbor search. More...

 
typedef DefeatistKNN< tree::SPTreeSpillKNN
 The SpillKNN class is the k-nearest-neighbors method considering defeatist search on SPTree. More...

 

Enumerations

enum  NeighborSearchMode
{
  NAIVE_MODE
,
  SINGLE_TREE_MODE
,
  DUAL_TREE_MODE
,
  GREEDY_SINGLE_TREE_MODE

}
 NeighborSearchMode represents the different neighbor search modes available. More...

 

Functions

void Unmap (const arma::Mat< size_t > &neighbors, const arma::mat &distances, const std::vector< size_t > &referenceMap, const std::vector< size_t > &queryMap, arma::Mat< size_t > &neighborsOut, arma::mat &distancesOut, const bool squareRoot=false)
 Assuming that the datasets have been mapped using the referenceMap and the queryMap (such as during kd-tree construction), unmap the columns of the distances and neighbors matrices into neighborsOut and distancesOut, and also unmap the entries in each row of neighbors. More...

 
void Unmap (const arma::Mat< size_t > &neighbors, const arma::mat &distances, const std::vector< size_t > &referenceMap, arma::Mat< size_t > &neighborsOut, arma::mat &distancesOut, const bool squareRoot=false)
 Assuming that the datasets have been mapped using referenceMap (such as during kd-tree construction), unmap the columns of the distances and neighbors matrices into neighborsOut and distancesOut, and also unmap the entries in each row of neighbors. More...

 

Detailed Description

Neighbor-search routines.

These include all-nearest-neighbors and all-furthest-neighbors searches.

Typedef Documentation

◆ AllkFN

It returns L2 distances (Euclidean distances) for each of the k furthest neighbors. This typedef will be removed in mlpack 3.0.0; use the KFN typedef instead.

Definition at line 83 of file typedef.hpp.

◆ AllkNN

It returns L2 distances (Euclidean distances) for each of the k nearest neighbors. This typedef will be removed in mlpack 3.0.0; use the KNN typedef instead.

Definition at line 75 of file typedef.hpp.

◆ AllkRAFN

It returns L2 distances for each of the k rank-approximate farthest-neighbors.

The approximation is controlled with two parameters (see allkrann_main.cpp) which can be specified at search time. So the tree building is done only once while the search can be performed multiple times with different approximation levels.

This typedef will be removed in mlpack 3.0.0; use the KRANN typedef instead.

Definition at line 76 of file ra_typedef.hpp.

◆ AllkRANN

It returns L2 distances for each of the k rank-approximate nearest-neighbors.

The approximation is controlled with two parameters (see allkrann_main.cpp) which can be specified at search time. So the tree building is done only once while the search can be performed multiple times with different approximation levels.

This typedef will be removed in mlpack 3.0.0; use the KRANN typedef instead.

Definition at line 61 of file ra_typedef.hpp.

◆ DefeatistKNN

template
<
template
<
typename
TreeMetricType
,
typename
TreeStatType
,
typename
TreeMatType
>
class
TreeType
=
tree::SPTree
>
using mlpack::neighbor::DefeatistKNN = typedef NeighborSearch< NearestNeighborSort, metric::EuclideanDistance, arma::mat, TreeType, TreeType<metric::EuclideanDistance, NeighborSearchStat<NearestNeighborSort>, arma::mat>::template DefeatistDualTreeTraverser, TreeType<metric::EuclideanDistance, NeighborSearchStat<NearestNeighborSort>, arma::mat>::template DefeatistSingleTreeTraverser>

The DefeatistKNN class is the k-nearest-neighbors method considering defeatist search.

It returns L2 distances (Euclidean distances) for each of the k nearest neighbors found.

Template Parameters
TreeTypeThe tree type to use; must adhere to the TreeType API, and implement Defeatist Traversers.

Definition at line 60 of file typedef.hpp.

◆ KFN

The KFN class is the k-furthest-neighbors method.

It returns L2 distances (Euclidean distances) for each of the k furthest neighbors.

Definition at line 38 of file typedef.hpp.

◆ KNN

The KNN class is the k-nearest-neighbors method.

It returns L2 distances (Euclidean distances) for each of the k nearest neighbors.

Definition at line 32 of file typedef.hpp.

◆ KRAFN

The KRAFN class is the k-rank-approximate-farthest-neighbors method.

It returns L2 distances for each of the k rank-approximate farthest-neighbors.

The approximation is controlled with two parameters (see allkrann_main.cpp) which can be specified at search time. So the tree building is done only once while the search can be performed multiple times with different approximation levels.

Definition at line 47 of file ra_typedef.hpp.

◆ KRANN

The KRANN class is the k-rank-approximate-nearest-neighbors method.

It returns L2 distances for each of the k rank-approximate nearest-neighbors.

The approximation is controlled with two parameters (see allkrann_main.cpp) which can be specified at search time. So the tree building is done only once while the search can be performed multiple times with different approximation levels.

Definition at line 36 of file ra_typedef.hpp.

◆ NSType

template
<
typename
SortPolicy
,
template
<
typename
TreeMetricType
,
typename
TreeStatType
,
typename
TreeMatType
>
class
TreeType
>
using mlpack::neighbor::NSType = typedef NeighborSearch<SortPolicy, metric::EuclideanDistance, arma::mat, TreeType, TreeType<metric::EuclideanDistance, NeighborSearchStat<SortPolicy>, arma::mat>::template DualTreeTraverser>

Alias template for euclidean neighbor search.

Definition at line 42 of file ns_model.hpp.

◆ SpillKNN

The SpillKNN class is the k-nearest-neighbors method considering defeatist search on SPTree.

It returns L2 distances (Euclidean distances) for each of the k nearest neighbors found.

Definition at line 67 of file typedef.hpp.

Enumeration Type Documentation

◆ NeighborSearchMode

NeighborSearchMode represents the different neighbor search modes available.

Enumerator
NAIVE_MODE 
SINGLE_TREE_MODE 
DUAL_TREE_MODE 
GREEDY_SINGLE_TREE_MODE 

Definition at line 38 of file neighbor_search.hpp.

Function Documentation

◆ Unmap() [1/2]

void mlpack::neighbor::Unmap ( const arma::Mat< size_t > &  neighbors,
const arma::mat &  distances,
const std::vector< size_t > &  referenceMap,
const std::vector< size_t > &  queryMap,
arma::Mat< size_t > &  neighborsOut,
arma::mat &  distancesOut,
const bool  squareRoot = false 
)

Assuming that the datasets have been mapped using the referenceMap and the queryMap (such as during kd-tree construction), unmap the columns of the distances and neighbors matrices into neighborsOut and distancesOut, and also unmap the entries in each row of neighbors.

This is useful for the dual-tree case.

Parameters
neighborsMatrix of neighbors resulting from neighbor search.
distancesMatrix of distances resulting from neighbor search.
referenceMapMapping of reference set to old points.
queryMapMapping of query set to old points.
neighborsOutMatrix to store unmapped neighbors into.
distancesOutMatrix to store unmapped distances into.
squareRootIf true, take the square root of the distances.

◆ Unmap() [2/2]

void mlpack::neighbor::Unmap ( const arma::Mat< size_t > &  neighbors,
const arma::mat &  distances,
const std::vector< size_t > &  referenceMap,
arma::Mat< size_t > &  neighborsOut,
arma::mat &  distancesOut,
const bool  squareRoot = false 
)

Assuming that the datasets have been mapped using referenceMap (such as during kd-tree construction), unmap the columns of the distances and neighbors matrices into neighborsOut and distancesOut, and also unmap the entries in each row of neighbors.

This is useful for the single-tree case.

Parameters
neighborsMatrix of neighbors resulting from neighbor search.
distancesMatrix of distances resulting from neighbor search.
referenceMapMapping of reference set to old points.
neighborsOutMatrix to store unmapped neighbors into.
distancesOutMatrix to store unmapped distances into.
squareRootIf true, take the square root of the distances.