The RASearch class: This class provides a generic manner to perform rank-approximate search via random-sampling. More...
Public Member Functions | |
RASearch (TreeType *referenceTree, const typename TreeType::Mat &referenceSet, const bool singleMode=false, const MetricType metric=MetricType()) | |
Initialize the RASearch object with the given reference dataset and pre-constructed tree. | |
RASearch (TreeType *referenceTree, TreeType *queryTree, const typename TreeType::Mat &referenceSet, const typename TreeType::Mat &querySet, const bool singleMode=false, const MetricType metric=MetricType()) | |
Initialize the RASearch object with the given datasets and pre-constructed trees. | |
RASearch (const typename TreeType::Mat &referenceSet, const bool naive=false, const bool singleMode=false, const size_t leafSize=20, const MetricType metric=MetricType()) | |
Initialize the RASearch object, passing only one dataset, which is used as both the query and the reference dataset. | |
RASearch (const typename TreeType::Mat &referenceSet, const typename TreeType::Mat &querySet, const bool naive=false, const bool singleMode=false, const size_t leafSize=20, const MetricType metric=MetricType()) | |
Initialize the RASearch object, passing both a query and reference dataset. | |
~RASearch () | |
Delete the RASearch object. | |
void | ResetQueryTree () |
This function recursively resets the RAQueryStat of the queryTree to set 'bound' to WorstDistance and the 'numSamplesMade' to 0. | |
void | Search (const size_t k, arma::Mat< size_t > &resultingNeighbors, arma::mat &distances, const double tau=5, const double alpha=0.95, const bool sampleAtLeaves=false, const bool firstLeafExact=false, const size_t singleSampleLimit=20) |
Compute the rank approximate nearest neighbors and store the output in the given matrices. | |
Private Member Functions | |
void | ResetRAQueryStat (TreeType *treeNode) |
Private Attributes | |
MetricType | metric |
Instantiation of kernel. | |
bool | naive |
Indicates if naive random sampling on the set is being used. | |
size_t | numberOfPrunes |
Total number of pruned nodes during the neighbor search. | |
std::vector< size_t > | oldFromNewQueries |
Permutations of query points during tree building. | |
std::vector< size_t > | oldFromNewReferences |
Permutations of reference points during tree building. | |
bool | ownQueryTree |
Indicates if we should free the query tree at deletion time. | |
bool | ownReferenceTree |
Indicates if we should free the reference tree at deletion time. | |
arma::mat | queryCopy |
Copy of query dataset (if we need it, because tree building modifies it). | |
const arma::mat & | querySet |
Query dataset (may not be given). | |
TreeType * | queryTree |
Pointer to the root of the query tree (might not exist). | |
arma::mat | referenceCopy |
Copy of reference dataset (if we need it, because tree building modifies it). | |
const arma::mat & | referenceSet |
Reference dataset. | |
TreeType * | referenceTree |
Pointer to the root of the reference tree. | |
bool | singleMode |
Indicates if single-tree search is being used (opposed to dual-tree). |
The RASearch class: This class provides a generic manner to perform rank-approximate search via random-sampling.
If the 'naive' option is chosen, this rank-approximate search will be done by randomly sampled from the whole set. If the 'naive' option is not chosen, the sampling is done in a stratified manner in the tree as mentioned in the algorithms in Figure 2 of the following paper:
{ram2009rank, title={{Rank-Approximate Nearest Neighbor Search: Retaining Meaning and Speed in High Dimensions}}, author={{Ram, P. and Lee, D. and Ouyang, H. and Gray, A. G.}}, booktitle={{Advances of Neural Information Processing Systems}}, year={2009} }
SortPolicy | The sort policy for distances; see NearestNeighborSort. | |
MetricType | The metric to use for computation. | |
TreeType | The tree type to use. |
Definition at line 117 of file ra_search.hpp.
mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::RASearch | ( | const typename TreeType::Mat & | referenceSet, | |
const typename TreeType::Mat & | querySet, | |||
const bool | naive = false , |
|||
const bool | singleMode = false , |
|||
const size_t | leafSize = 20 , |
|||
const MetricType | metric = MetricType() | |||
) |
Initialize the RASearch object, passing both a query and reference dataset.
Optionally, perform the computation in naive mode or single-tree mode, and set the leaf size used for tree-building. An initialized distance metric can be given, for cases where the metric has internal data (i.e. the distance::MahalanobisDistance class).
This method will copy the matrices to internal copies, which are rearranged during tree-building. You can avoid this extra copy by pre-constructing the trees and passing them using a diferent constructor.
referenceSet | Set of reference points. | |
querySet | Set of query points. | |
naive | If true, the rank-approximate search will be performed by directly sampling the whole set instead of using the stratified sampling on the tree. | |
singleMode | If true, single-tree search will be used (as opposed to dual-tree search). | |
leafSize | Leaf size for tree construction (ignored if tree is given). | |
metric | An optional instance of the MetricType class. |
mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::RASearch | ( | const typename TreeType::Mat & | referenceSet, | |
const bool | naive = false , |
|||
const bool | singleMode = false , |
|||
const size_t | leafSize = 20 , |
|||
const MetricType | metric = MetricType() | |||
) |
Initialize the RASearch object, passing only one dataset, which is used as both the query and the reference dataset.
Optionally, perform the computation in naive mode or single-tree mode, and set the leaf size used for tree-building. An initialized distance metric can be given, for cases where the metric has internal data (i.e. the distance::MahalanobisDistance class).
If naive mode is being used and a pre-built tree is given, it may not work: naive mode operates by building a one-node tree (the root node holds all the points). If that condition is not satisfied with the pre-built tree, then naive mode will not work.
referenceSet | Set of reference points. | |
naive | If true, the rank-approximate search will be performed by directly sampling the whole set instead of using the stratified sampling on the tree. | |
singleMode | If true, single-tree search will be used (as opposed to dual-tree search). | |
leafSize | Leaf size for tree construction (ignored if tree is given). | |
metric | An optional instance of the MetricType class. |
mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::RASearch | ( | TreeType * | referenceTree, | |
TreeType * | queryTree, | |||
const typename TreeType::Mat & | referenceSet, | |||
const typename TreeType::Mat & | querySet, | |||
const bool | singleMode = false , |
|||
const MetricType | metric = MetricType() | |||
) |
Initialize the RASearch object with the given datasets and pre-constructed trees.
It is assumed that the points in referenceSet and querySet correspond to the points in referenceTree and queryTree, respectively. Optionally, choose to use single-tree mode. Naive mode is not available as an option for this constructor; instead, to run naive computation, construct a tree with all of the points in one leaf (i.e. leafSize = number of points). Additionally, an instantiated distance metric can be given, for cases where the distance metric holds data.
There is no copying of the data matrices in this constructor (because tree-building is not necessary), so this is the constructor to use when copies absolutely must be avoided.
referenceTree | Pre-built tree for reference points. | |
queryTree | Pre-built tree for query points. | |
referenceSet | Set of reference points corresponding to referenceTree. | |
querySet | Set of query points corresponding to queryTree. | |
singleMode | Whether single-tree computation should be used (as opposed to dual-tree computation). | |
metric | Instantiated distance metric. |
mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::RASearch | ( | TreeType * | referenceTree, | |
const typename TreeType::Mat & | referenceSet, | |||
const bool | singleMode = false , |
|||
const MetricType | metric = MetricType() | |||
) |
Initialize the RASearch object with the given reference dataset and pre-constructed tree.
It is assumed that the points in referenceSet correspond to the points in referenceTree. Optionally, choose to use single-tree mode. Naive mode is not available as an option for this constructor; instead, to run naive computation, construct a tree with all the points in one leaf (i.e. leafSize = number of points). Additionally, an instantiated distance metric can be given, for the case where the distance metric holds data.
There is no copying of the data matrices in this constructor (because tree-building is not necessary), so this is the constructor to use when copies absolutely must be avoided.
mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::~RASearch | ( | ) |
void mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::ResetQueryTree | ( | ) |
This function recursively resets the RAQueryStat of the queryTree to set 'bound' to WorstDistance and the 'numSamplesMade' to 0.
This allows a user to perform multiple searches on the same pair of trees, possibly with different levels of approximation without requiring to build a new pair of trees for every new (approximate) search.
void mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::ResetRAQueryStat | ( | TreeType * | treeNode | ) | [private] |
treeNode | The node of the tree whose RAQueryStat is reset and whose children are to be explored recursively. |
void mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::Search | ( | const size_t | k, | |
arma::Mat< size_t > & | resultingNeighbors, | |||
arma::mat & | distances, | |||
const double | tau = 5 , |
|||
const double | alpha = 0.95 , |
|||
const bool | sampleAtLeaves = false , |
|||
const bool | firstLeafExact = false , |
|||
const size_t | singleSampleLimit = 20 | |||
) |
Compute the rank approximate 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.
Note that tau, the rank-approximation parameter, specifies that we are looking for k neighbors with probability alpha of being in the top tau percent of nearest neighbors. So, as an example, if our dataset has 1000 points, and we want 5 nearest neighbors with 95% probability of being in the top 5% of nearest neighbors (or, the top 50 nearest neighbors), we set k = 5, tau = 5, and alpha = 0.95.
The method will fail (and issue a failure message) if the value of tau is too low: tau must be set such that the number of points in the corresponding percentile of the data is greater than k. Thus, if we choose tau = 0.1 with a dataset of 1000 points and k = 5, then we are attempting to choose 5 nearest neighbors out of the closest 1 point -- this is invalid.
k | Number of neighbors to search for. | |
resultingNeighbors | Matrix storing lists of neighbors for each query point. | |
distances | Matrix storing distances of neighbors for each query point. | |
tau | The rank-approximation in percentile of the data. The default value is 5%. | |
alpha | The desired success probability. The default value is 0.95. | |
sampleAtLeaves | Sample at leaves for faster but less accurate computation. This defaults to 'false'. | |
firstLeafExact | Traverse to the first leaf without approximation. This can ensure that the query definitely finds its (near) duplicate if there exists one. This defaults to 'false' for now. | |
singleSampleLimit | The limit on the largest node that can be approximated by sampling. This defaults to 20. |
MetricType mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::metric [private] |
Instantiation of kernel.
Definition at line 332 of file ra_search.hpp.
bool mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::naive [private] |
Indicates if naive random sampling on the set is being used.
Definition at line 327 of file ra_search.hpp.
size_t mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::numberOfPrunes [private] |
Total number of pruned nodes during the neighbor search.
Definition at line 340 of file ra_search.hpp.
std::vector<size_t> mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::oldFromNewQueries [private] |
Permutations of query points during tree building.
Definition at line 337 of file ra_search.hpp.
std::vector<size_t> mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::oldFromNewReferences [private] |
Permutations of reference points during tree building.
Definition at line 335 of file ra_search.hpp.
bool mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::ownQueryTree [private] |
Indicates if we should free the query tree at deletion time.
Definition at line 324 of file ra_search.hpp.
bool mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::ownReferenceTree [private] |
Indicates if we should free the reference tree at deletion time.
Definition at line 322 of file ra_search.hpp.
arma::mat mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::queryCopy [private] |
Copy of query dataset (if we need it, because tree building modifies it).
Definition at line 309 of file ra_search.hpp.
const arma::mat& mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::querySet [private] |
Query dataset (may not be given).
Definition at line 314 of file ra_search.hpp.
TreeType* mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::queryTree [private] |
Pointer to the root of the query tree (might not exist).
Definition at line 319 of file ra_search.hpp.
arma::mat mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::referenceCopy [private] |
Copy of reference dataset (if we need it, because tree building modifies it).
Definition at line 307 of file ra_search.hpp.
const arma::mat& mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::referenceSet [private] |
Reference dataset.
Definition at line 312 of file ra_search.hpp.
TreeType* mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::referenceTree [private] |
Pointer to the root of the reference tree.
Definition at line 317 of file ra_search.hpp.
bool mlpack::neighbor::RASearch< SortPolicy, MetricType, TreeType >::singleMode [private] |
Indicates if single-tree search is being used (opposed to dual-tree).
Definition at line 329 of file ra_search.hpp.