de.lmu.ifi.dbs.elki.distance.distancefunction.correlation
Class PearsonCorrelationDistanceFunction<V extends NumberVector<V,N>,N extends Number>
java.lang.Object
de.lmu.ifi.dbs.elki.logging.AbstractLoggable
de.lmu.ifi.dbs.elki.distance.AbstractMeasurementFunction<O,D>
de.lmu.ifi.dbs.elki.distance.distancefunction.AbstractDistanceFunction<V,DoubleDistance>
de.lmu.ifi.dbs.elki.distance.distancefunction.correlation.PearsonCorrelationDistanceFunction<V,N>
- Type Parameters:
V
- the type of FeatureVector to compute the distances in betweenN
- the type of Number of the attributes of vectors of type V
- All Implemented Interfaces:
- DistanceFunction<V,DoubleDistance>, MeasurementFunction<V,DoubleDistance>, Parameterizable
public class PearsonCorrelationDistanceFunction<V extends NumberVector<V,N>,N extends Number>
- extends AbstractDistanceFunction<V,DoubleDistance>
- implements Parameterizable
Pearson correlation distance function for feature vectors.
The Pearson correlation distance is computed from the Pearson correlation coefficient r
as: 1-r
.
Hence, possible values of this distance are between 0 and 2.
The distance between two vectors will be low (near 0), if their attribute values are dimension-wise strictly positively correlated,
it will be high (near 2), if their attribute values are dimension-wise strictly negatively correlated.
For Features with uncorrelated attributes, the distance value will be intermediate (around 1).
- Author:
- Arthur Zimek
Method Summary |
DoubleDistance |
distance(V v1,
V v2)
Computes the Pearson correlation distance for two given feature vectors. |
Methods inherited from class java.lang.Object |
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait |
PearsonCorrelationDistanceFunction
public PearsonCorrelationDistanceFunction()
- Provides a PearsonCorrelationDistanceFunction.
distance
public DoubleDistance distance(V v1,
V v2)
- Computes the Pearson correlation distance for two given feature vectors.
The Pearson correlation distance is computed from the Pearson correlation coefficient
r
as: 1-r
.
Hence, possible values of this distance are between 0 and 2.
- Specified by:
distance
in interface DistanceFunction<V extends NumberVector<V,N>,DoubleDistance>
- Parameters:
v1
- first feature vectorv2
- second feature vector
- Returns:
- the Pearson correlation distance for two given feature vectors v1 and v2