org.deeplearning4j.clustering.vptree

Class VPTree

java.lang.Object

org.deeplearning4j.clustering.vptree.VPTree

public class VPTree
extends Object

Vantage point tree implementation

Author:

Adam Gibson

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	VPTree.Node

Field Summary

Fields

Modifier and Type	Field and Description
static String	EUCLIDEAN

Constructor Summary

Constructors

Constructor and Description
VPTree(org.nd4j.linalg.api.ndarray.INDArray items)
VPTree(org.nd4j.linalg.api.ndarray.INDArray items, String similarityFunction)
VPTree(org.nd4j.linalg.api.ndarray.INDArray items, String similarityFunction, boolean invert)
VPTree(List<DataPoint> items)
VPTree(List<DataPoint> items, CounterMap<DataPoint,DataPoint> distances)
VPTree(List<DataPoint> items, CounterMap<DataPoint,DataPoint> distances, String similarityFunction)
VPTree(List<DataPoint> items, CounterMap<DataPoint,DataPoint> distances, String similarityFunction, boolean invert)
VPTree(List<DataPoint> items, String similarityFunction)
VPTree(List<DataPoint> items, String similarityFunction, boolean invert)

Method Summary

Modifier and Type	Method and Description
static org.nd4j.linalg.api.ndarray.INDArray	buildFromData(List<DataPoint> data)
CounterMap<DataPoint,DataPoint>	getDistances()
List<DataPoint>	getItems()
void	search(DataPoint target, int k, List<DataPoint> results, List<Double> distances)
void	search(VPTree.Node node, DataPoint target, int k, PriorityQueue<HeapItem> pq)
void	setDistances(CounterMap<DataPoint,DataPoint> distances)
void	setItems(List<DataPoint> items)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

EUCLIDEAN

public static final String EUCLIDEAN

See Also:

Constant Field Values

Constructor Detail

VPTree

public VPTree(org.nd4j.linalg.api.ndarray.INDArray items,
              String similarityFunction,
              boolean invert)

Parameters:

items - the items to use

similarityFunction - the similiarity function to use

invert - whether to invert the distance (similarity functions have different min/max objectives)

VPTree

public VPTree(List<DataPoint> items,
              CounterMap<DataPoint,DataPoint> distances,
              String similarityFunction,
              boolean invert)

Parameters:

items - the items to use

distances - the distances

similarityFunction - the similarity function to use

invert - whether to invert the metric (different optimization objective)

VPTree

public VPTree(List<DataPoint> items,
              String similarityFunction,
              boolean invert)

VPTree

public VPTree(org.nd4j.linalg.api.ndarray.INDArray items,
              String similarityFunction)

VPTree

public VPTree(List<DataPoint> items,
              CounterMap<DataPoint,DataPoint> distances,
              String similarityFunction)

VPTree

public VPTree(List<DataPoint> items,
              String similarityFunction)

VPTree

public VPTree(org.nd4j.linalg.api.ndarray.INDArray items)

VPTree

public VPTree(List<DataPoint> items,
              CounterMap<DataPoint,DataPoint> distances)

VPTree

public VPTree(List<DataPoint> items)

Method Detail

buildFromData

public static org.nd4j.linalg.api.ndarray.INDArray buildFromData(List<DataPoint> data)

getItems

public List<DataPoint> getItems()

setItems

public void setItems(List<DataPoint> items)

search

public void search(DataPoint target,
                   int k,
                   List<DataPoint> results,
                   List<Double> distances)

search

public void search(VPTree.Node node,
                   DataPoint target,
                   int k,
                   PriorityQueue<HeapItem> pq)

getDistances

public CounterMap<DataPoint,DataPoint> getDistances()

setDistances

public void setDistances(CounterMap<DataPoint,DataPoint> distances)