MLPACK  1.0.10
Public Member Functions | Private Member Functions | Private Attributes | List of all members
mlpack::neighbor::LSHSearch< SortPolicy > Class Template Reference

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...

Public Member Functions

 LSHSearch (const arma::mat &referenceSet, const arma::mat &querySet, const size_t numProj, const size_t numTables, const double hashWidth=0.0, const size_t secondHashSize=99901, const size_t bucketSize=500)
 This function initializes the LSH class. More...
 
 LSHSearch (const arma::mat &referenceSet, const size_t numProj, const size_t numTables, const double hashWidth=0.0, const size_t secondHashSize=99901, const size_t bucketSize=500)
 This function initializes the LSH class. More...
 
void Search (const size_t k, arma::Mat< size_t > &resultingNeighbors, arma::mat &distances, const size_t numTablesToSearch=0)
 Compute the nearest neighbors and store the output in the given matrices. More...
 
std::string ToString () const
 

Private Member Functions

double BaseCase (const size_t queryIndex, const size_t referenceIndex)
 This is a helper function that computes the distance of the query to the neighbor candidates and appropriately stores the best 'k' candidates. More...
 
void BuildHash ()
 This function builds a hash table with two levels of hashing as presented in the paper. More...
 
void InsertNeighbor (const size_t queryIndex, const size_t pos, const size_t neighbor, const double distance)
 This is a helper function that efficiently inserts better neighbor candidates into an existing set of neighbor candidates. More...
 
void ReturnIndicesFromTable (const size_t queryIndex, arma::uvec &referenceIndices, size_t numTablesToSearch)
 This function takes a query and hashes it into each of the hash tables to get keys for the query and then the key is hashed to a bucket of the second hash table and all the points (if any) in those buckets are collected as the potential neighbor candidates. More...
 

Private Attributes

arma::Col< size_t > bucketContentSize
 The number of elements present in each hash bucket; should be secondHashSize. More...
 
arma::Col< size_t > bucketRowInHashTable
 For a particular hash value, points to the row in secondHashTable corresponding to this value. More...
 
const size_t bucketSize
 The bucket size of the second hash. More...
 
arma::mat * distancePtr
 The pointer to the nearest neighbor distances. More...
 
double hashWidth
 The hash width. More...
 
metric::SquaredEuclideanDistance metric
 Instantiation of the metric. More...
 
arma::Mat< size_t > * neighborPtr
 The pointer to the nearest neighbor indices. More...
 
const size_t numProj
 The number of projections. More...
 
const size_t numTables
 The number of hash tables. More...
 
arma::mat offsets
 The list of the offset 'b' for each of the projection for each table. More...
 
std::vector< arma::mat > projections
 The std::vector containing the projection matrix of each table. More...
 
const arma::mat & querySet
 Query dataset (may not be given). More...
 
const arma::mat & referenceSet
 Reference dataset. More...
 
const size_t secondHashSize
 The big prime representing the size of the second hash. More...
 
arma::Mat< size_t > secondHashTable
 The final hash table; should be (< secondHashSize) x bucketSize. More...
 
arma::vec secondHashWeights
 The weights of the second hash. More...
 

Detailed Description

template<typename SortPolicy = NearestNeighborSort>
class mlpack::neighbor::LSHSearch< SortPolicy >

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.

Template Parameters
SortPolicyThe sort policy for distances; see NearestNeighborSort.

Definition at line 59 of file lsh_search.hpp.

Constructor & Destructor Documentation

template<typename SortPolicy = NearestNeighborSort>
mlpack::neighbor::LSHSearch< SortPolicy >::LSHSearch ( const arma::mat &  referenceSet,
const arma::mat &  querySet,
const size_t  numProj,
const size_t  numTables,
const double  hashWidth = 0.0,
const size_t  secondHashSize = 99901,
const size_t  bucketSize = 500 
)

This function initializes the LSH class.

It builds the hash on the reference set with 2-stable distributions. See the individual functions performing the hashing for details on how the hashing is done.

Parameters
referenceSetSet of reference points.
querySetSet of query points.
numProjNumber of projections in each hash table (anything between 10-50 might be a decent choice).
numTablesTotal number of hash tables (anything between 10-20 should suffice).
hashWidthThe width of hash for every table. If 0 (the default) is provided, then the hash width is automatically obtained by computing the average pairwise distance of 25 pairs. This should be a reasonable upper bound on the nearest-neighbor distance in general.
secondHashSizeThe size of the second hash table. This should be a large prime number.
bucketSizeThe size of the bucket in the second hash table. This is the maximum number of points that can be hashed into single bucket. Default values are already provided here.
template<typename SortPolicy = NearestNeighborSort>
mlpack::neighbor::LSHSearch< SortPolicy >::LSHSearch ( const arma::mat &  referenceSet,
const size_t  numProj,
const size_t  numTables,
const double  hashWidth = 0.0,
const size_t  secondHashSize = 99901,
const size_t  bucketSize = 500 
)

This function initializes the LSH class.

It builds the hash on the reference set with 2-stable distributions. See the individual functions performing the hashing for details on how the hashing is done.

Parameters
referenceSetSet of reference points and the set of queries.
numProjNumber of projections in each hash table (anything between 10-50 might be a decent choice).
numTablesTotal number of hash tables (anything between 10-20 should suffice).
hashWidthThe width of hash for every table. If 0 (the default) is provided, then the hash width is automatically obtained by computing the average pairwise distance of 25 pairs. This should be a reasonable upper bound on the nearest-neighbor distance in general.
secondHashSizeThe size of the second hash table. This should be a large prime number.
bucketSizeThe size of the bucket in the second hash table. This is the maximum number of points that can be hashed into single bucket. Default values are already provided here.

Member Function Documentation

template<typename SortPolicy = NearestNeighborSort>
double mlpack::neighbor::LSHSearch< SortPolicy >::BaseCase ( const size_t  queryIndex,
const size_t  referenceIndex 
)
private

This is a helper function that computes the distance of the query to the neighbor candidates and appropriately stores the best 'k' candidates.

Parameters
queryIndexThe index of the query in question
referenceIndexThe index of the neighbor candidate in question
template<typename SortPolicy = NearestNeighborSort>
void mlpack::neighbor::LSHSearch< SortPolicy >::BuildHash ( )
private

This function builds a hash table with two levels of hashing as presented in the paper.

This function first hashes the points with 'numProj' random projections to a single hash table creating (key, point ID) pairs where the key is a 'numProj'-dimensional integer vector.

Then each key in this hash table is hashed into a second hash table using a standard hash.

This function does not have any parameters and relies on parameters which are private members of this class, intialized during the class intialization.

template<typename SortPolicy = NearestNeighborSort>
void mlpack::neighbor::LSHSearch< SortPolicy >::InsertNeighbor ( const size_t  queryIndex,
const size_t  pos,
const size_t  neighbor,
const double  distance 
)
private

This is a helper function that efficiently inserts better neighbor candidates into an existing set of neighbor candidates.

This function is only called by the 'BaseCase' function.

Parameters
queryIndexThis is the index of the query being processed currently
posThe position of the neighbor candidate in the current list of neighbor candidates.
neighborThe neighbor candidate that is being inserted into the list of the best 'k' candidates for the query in question.
distanceThe distance of the query to the neighbor candidate.
template<typename SortPolicy = NearestNeighborSort>
void mlpack::neighbor::LSHSearch< SortPolicy >::ReturnIndicesFromTable ( const size_t  queryIndex,
arma::uvec &  referenceIndices,
size_t  numTablesToSearch 
)
private

This function takes a query and hashes it into each of the hash tables to get keys for the query and then the key is hashed to a bucket of the second hash table and all the points (if any) in those buckets are collected as the potential neighbor candidates.

Parameters
queryIndexThe index of the query currently being processed.
referenceIndicesThe list of neighbor candidates obtained from hashing the query into all the hash tables and eventually into multiple buckets of the second hash table.
template<typename SortPolicy = NearestNeighborSort>
void mlpack::neighbor::LSHSearch< SortPolicy >::Search ( const size_t  k,
arma::Mat< size_t > &  resultingNeighbors,
arma::mat &  distances,
const size_t  numTablesToSearch = 0 
)

Compute the nearest neighbors and store the output in the given matrices.

The matrices will be set to the size of n columns by k rows, where n is the number of points in the query dataset and k is the number of neighbors being searched for.

Parameters
kNumber of neighbors to search for.
resultingNeighborsMatrix storing lists of neighbors for each query point.
distancesMatrix storing distances of neighbors for each query point.
numTablesToSearchThis parameter allows the user to have control over the number of hash tables to be searched. This allows the user to pick the number of tables it can afford for the time available without having to build hashing for every table size. By default, this is set to zero in which case all tables are considered.
template<typename SortPolicy = NearestNeighborSort>
std::string mlpack::neighbor::LSHSearch< SortPolicy >::ToString ( ) const

Member Data Documentation

template<typename SortPolicy = NearestNeighborSort>
arma::Col<size_t> mlpack::neighbor::LSHSearch< SortPolicy >::bucketContentSize
private

The number of elements present in each hash bucket; should be secondHashSize.

Definition at line 237 of file lsh_search.hpp.

template<typename SortPolicy = NearestNeighborSort>
arma::Col<size_t> mlpack::neighbor::LSHSearch< SortPolicy >::bucketRowInHashTable
private

For a particular hash value, points to the row in secondHashTable corresponding to this value.

Should be secondHashSize.

Definition at line 241 of file lsh_search.hpp.

template<typename SortPolicy = NearestNeighborSort>
const size_t mlpack::neighbor::LSHSearch< SortPolicy >::bucketSize
private

The bucket size of the second hash.

Definition at line 227 of file lsh_search.hpp.

template<typename SortPolicy = NearestNeighborSort>
arma::mat* mlpack::neighbor::LSHSearch< SortPolicy >::distancePtr
private

The pointer to the nearest neighbor distances.

Definition at line 244 of file lsh_search.hpp.

template<typename SortPolicy = NearestNeighborSort>
double mlpack::neighbor::LSHSearch< SortPolicy >::hashWidth
private

The hash width.

Definition at line 218 of file lsh_search.hpp.

template<typename SortPolicy = NearestNeighborSort>
metric::SquaredEuclideanDistance mlpack::neighbor::LSHSearch< SortPolicy >::metric
private

Instantiation of the metric.

Definition at line 230 of file lsh_search.hpp.

template<typename SortPolicy = NearestNeighborSort>
arma::Mat<size_t>* mlpack::neighbor::LSHSearch< SortPolicy >::neighborPtr
private

The pointer to the nearest neighbor indices.

Definition at line 247 of file lsh_search.hpp.

template<typename SortPolicy = NearestNeighborSort>
const size_t mlpack::neighbor::LSHSearch< SortPolicy >::numProj
private

The number of projections.

Definition at line 206 of file lsh_search.hpp.

template<typename SortPolicy = NearestNeighborSort>
const size_t mlpack::neighbor::LSHSearch< SortPolicy >::numTables
private

The number of hash tables.

Definition at line 209 of file lsh_search.hpp.

template<typename SortPolicy = NearestNeighborSort>
arma::mat mlpack::neighbor::LSHSearch< SortPolicy >::offsets
private

The list of the offset 'b' for each of the projection for each table.

Definition at line 215 of file lsh_search.hpp.

template<typename SortPolicy = NearestNeighborSort>
std::vector<arma::mat> mlpack::neighbor::LSHSearch< SortPolicy >::projections
private

The std::vector containing the projection matrix of each table.

Definition at line 212 of file lsh_search.hpp.

template<typename SortPolicy = NearestNeighborSort>
const arma::mat& mlpack::neighbor::LSHSearch< SortPolicy >::querySet
private

Query dataset (may not be given).

Definition at line 203 of file lsh_search.hpp.

template<typename SortPolicy = NearestNeighborSort>
const arma::mat& mlpack::neighbor::LSHSearch< SortPolicy >::referenceSet
private

Reference dataset.

Definition at line 200 of file lsh_search.hpp.

template<typename SortPolicy = NearestNeighborSort>
const size_t mlpack::neighbor::LSHSearch< SortPolicy >::secondHashSize
private

The big prime representing the size of the second hash.

Definition at line 221 of file lsh_search.hpp.

template<typename SortPolicy = NearestNeighborSort>
arma::Mat<size_t> mlpack::neighbor::LSHSearch< SortPolicy >::secondHashTable
private

The final hash table; should be (< secondHashSize) x bucketSize.

Definition at line 233 of file lsh_search.hpp.

template<typename SortPolicy = NearestNeighborSort>
arma::vec mlpack::neighbor::LSHSearch< SortPolicy >::secondHashWeights
private

The weights of the second hash.

Definition at line 224 of file lsh_search.hpp.


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