LinearWeightedExtendedNeighborhood (ELKI: Environment for DeveLoping KDD-Applications Supported by Index-Structures)

java.lang.Object
- elki.outlier.spatial.neighborhood.weighted.LinearWeightedExtendedNeighborhood

All Implemented Interfaces:

WeightedNeighborSetPredicate
```
public class LinearWeightedExtendedNeighborhood
extends java.lang.Object
implements WeightedNeighborSetPredicate
```
Neighborhood obtained by computing the k-fold closure of an existing neighborhood. Objects are weighted linearly by their distance: the object itself has a weight of 1 and this decreases linearly to 1/(n+1) for the nth-step neighbors.
TODO: make actual weighting parameterizable?

Since:

0.4.0

Author:

Erich Schubert

Nested Class Summary

Nested Classes
Modifier and Type Class Description

static class LinearWeightedExtendedNeighborhood.Factory<O>
Factory class.

Field Summary

Fields
Modifier and Type Field Description

private NeighborSetPredicate inner
The data store to use

private int steps
The number of steps to extend to.

Constructor Summary

Constructors
Constructor Description

LinearWeightedExtendedNeighborhood(NeighborSetPredicate inner, int steps)
Constructor.

Method Summary

All Methods Instance Methods Concrete Methods
Modifier and Type	Method	Description
`private double`	`computeWeight(int tsteps)`	Compute the weight from the number of steps needed.
`java.util.Collection<DoubleDBIDPair>`	`getWeightedNeighbors(DBIDRef reference)`	Get the neighbors of a reference object for DBSCAN.

Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

- Field Detail
  - inner
```
private NeighborSetPredicate inner
```
    The data store to use
  - steps
```
private int steps
```
    The number of steps to extend to.
- Constructor Detail
  - LinearWeightedExtendedNeighborhood
```
public LinearWeightedExtendedNeighborhood(NeighborSetPredicate inner,
                                          int steps)
```
    Constructor.
    
    Parameters:
    
    inner - Inner neighborhood
    
    steps - Number of steps to expand
- Method Detail
  - computeWeight
```
private double computeWeight(int tsteps)
```
    Compute the weight from the number of steps needed.
    
    Parameters:
    
    tsteps - steps to target
    
    Returns:
    
    weight
  - getWeightedNeighbors
```
public java.util.Collection<DoubleDBIDPair> getWeightedNeighbors(DBIDRef reference)
```
    Description copied from interface: WeightedNeighborSetPredicate
    
    Get the neighbors of a reference object for DBSCAN.
    
    Specified by:
    
    getWeightedNeighbors in interface WeightedNeighborSetPredicate
    
    Parameters:
    
    reference - Reference object
    
    Returns:
    
    Weighted Neighborhood