org.openimaj.math.matrix.similarity.processor

Class MultidimensionalScaling

java.lang.Object

org.openimaj.math.matrix.similarity.processor.MultidimensionalScaling

All Implemented Interfaces:

SimilarityMatrixProcessor

public class MultidimensionalScaling
extends Object
implements SimilarityMatrixProcessor

Implementation of Multidimensional Scaling.

Implementation originally based around Toby Segaran's python code.

Author:

Jonathon Hare (jsh2@ecs.soton.ac.uk)

See Also:

"http://blog.kiwitobes.com/?p=44"

Field Summary

Fields

Modifier and Type	Field and Description
protected int	numIterations
protected List<IndependentPair<String,Point2d>>	points
protected double	rate
protected Random	rng

Constructor Summary

Constructors

Constructor and Description
MultidimensionalScaling() Default constructor.
MultidimensionalScaling(int numIterations, double rate) Construct MDS with the given maximum number of iterations and rate.
MultidimensionalScaling(int numIterations, double rate, Random rng) Construct MDS with the given maximum number of iterations, rate and random number generator.
MultidimensionalScaling(Random rng) Construct with the given random number generator and default learning rate at 0.01 and the maximum number of iterations to 1000.

Method Summary

Modifier and Type	Method and Description
Point2d	getPoint(String key) Get the predicted point for a specific element.
List<IndependentPair<String,Point2d>>	getPoints() Get a list of the 2-D coordinates learned by the MDS algorithm for each element in the input similarity matrix.
void	process(SimilarityMatrix matrix) Process the SimilarityMatrix, making changes inplace.
String	toString()

Methods inherited from class java.lang.Object

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

Field Detail

rng

protected Random rng

numIterations

protected int numIterations

rate

protected double rate

points

protected List<IndependentPair<String,Point2d>> points

Constructor Detail

MultidimensionalScaling

public MultidimensionalScaling()

Default constructor. Sets the learning rate at 0.01 and the maximum number of iterations to 1000.

MultidimensionalScaling

public MultidimensionalScaling(Random rng)

Construct with the given random number generator and default learning rate at 0.01 and the maximum number of iterations to 1000.

Parameters:

rng - the random number generator

MultidimensionalScaling

public MultidimensionalScaling(int numIterations,
                               double rate)

Construct MDS with the given maximum number of iterations and rate.

Parameters:

numIterations - number of iterations

rate - learning rate

MultidimensionalScaling

public MultidimensionalScaling(int numIterations,
                               double rate,
                               Random rng)

Construct MDS with the given maximum number of iterations, rate and random number generator.

Parameters:

numIterations - number of iterations

rate - learning rate

rng - the random number generator

Method Detail

process

public void process(SimilarityMatrix matrix)

Description copied from interface: SimilarityMatrixProcessor

Process the SimilarityMatrix, making changes inplace.

Specified by:

process in interface SimilarityMatrixProcessor

Parameters:

matrix - the matrix to process.

getPoints

public List<IndependentPair<String,Point2d>> getPoints()

Get a list of the 2-D coordinates learned by the MDS algorithm for each element in the input similarity matrix.

Returns:

list of <index, point>

getPoint

public Point2d getPoint(String key)

Get the predicted point for a specific element.

Parameters:

key - the element identifier

Returns:

the predicted point, or null if the key was not found.

toString

public String toString()

Overrides:

toString in class Object