com.aliasi.util

Interface Distance<E>

Type Parameters:

E - the type of objects over which distances are defined

All Known Implementing Classes:

EditDistance, EuclideanDistance, FixedWeightEditDistance, JaccardDistance, JaroWinklerDistance, MinkowskiDistance, TaxicabDistance, TfIdfDistance, TokenizedDistance, WeightedEditDistance

public interface Distance<E>

The Distance interface provides a general method for defining distances between two objects. Distance is a kind of dissimilarity measure, because the larger the distance between two objects, the less similar they are. The distance interface provides a single method distance(Object,Object) returning the distance between objects.

A proper distance is said to form a metric if it satisfies the following four properties:

• Positive: distance(x,y) >= 0
• Self Distance Zero: distance(x,x) = 0
• Symmetric: distance(x,y) = distance(y,x)
• Triangle Inequaltiy: distance(x,y) + distance(y,z) >= distance(x,z)

For example, the Euclidean distance between vectors is a proper metric.

 distance(x,y) = sqrt(Σi (x[i] * y[i])2)

as is the Manhattan metric, or taxicab distance:

 distance(x,y) = Σi abs(x[i] - y[i])

Cosine is also popular for vectors:

 distance(x,y) = dotProduct(x,y) / (length(x) * length(y))

A good introduction to distance may be found at:

• Wikipedia: Distance

Since:

LingPipe3.0

Version:

3.0

Author:

Bob Carpenter

Method Summary

Methods

Modifier and Type	Method and Description
double	distance(E e1, E e2) Returns the distance between the specified pair of objects.

Method Detail

distance

double distance(E e1,
              E e2)

Returns the distance between the specified pair of objects.

Parameters:

e1 - First object.

e2 - Second object.

Returns:

Distance between the two objects.