org.openimaj.ml.pca

Class FeatureVectorPCA

java.lang.Object

org.openimaj.math.matrix.algorithm.pca.PrincipalComponentAnalysis

org.openimaj.ml.pca.FeatureVectorPCA

public class FeatureVectorPCA
extends org.openimaj.math.matrix.algorithm.pca.PrincipalComponentAnalysis

Principal Components Analysis wrapper for FeatureVectors.

Author:

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

Nested Class Summary

Nested classes/interfaces inherited from class org.openimaj.math.matrix.algorithm.pca.PrincipalComponentAnalysis

org.openimaj.math.matrix.algorithm.pca.PrincipalComponentAnalysis.ComponentSelector, org.openimaj.math.matrix.algorithm.pca.PrincipalComponentAnalysis.EnergyThresholdComponentSelector, org.openimaj.math.matrix.algorithm.pca.PrincipalComponentAnalysis.NumberComponentSelector, org.openimaj.math.matrix.algorithm.pca.PrincipalComponentAnalysis.PercentageEnergyComponentSelector

Field Summary

Fields inherited from class org.openimaj.math.matrix.algorithm.pca.PrincipalComponentAnalysis

basis, eigenvalues, mean

Constructor Summary

Constructors

Constructor and Description
FeatureVectorPCA() Default constructor, using an SvdPrincipalComponentAnalysis.
FeatureVectorPCA(org.openimaj.math.matrix.algorithm.pca.PrincipalComponentAnalysis inner) Construct with the given PrincipalComponentAnalysis object.

Method Summary

Modifier and Type	Method and Description
DoubleFV	generate(DoubleFV scalings) Generate a new "observation" as a linear combination of the principal components (PC): mean + PC * scaling.
void	learnBasis(Collection<? extends FeatureVector> data) Learn the PCA basis of the given feature vectors.
void	learnBasis(double[][] data)
void	learnBasis(FeatureVector[] data) Learn the PCA basis of the given feature vectors.
protected void	learnBasisNorm(Jama.Matrix norm)
DoubleFV	project(FeatureVector vector) Project a vector by the basis.

Methods inherited from class org.openimaj.math.matrix.algorithm.pca.PrincipalComponentAnalysis

buildNormalisedDataMatrix, generate, getBasis, getCumulativeEnergies, getCumulativeEnergy, getEigenValue, getEigenValues, getEigenVectors, getMean, getPrincipalComponent, getStandardDeviations, getStandardDeviations, learnBasis, learnBasis, project, project, selectSubset, selectSubset, selectSubsetEnergyThreshold, selectSubsetPercentageEnergy, toString

Methods inherited from class java.lang.Object

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

Constructor Detail

FeatureVectorPCA

public FeatureVectorPCA()

Default constructor, using an SvdPrincipalComponentAnalysis.

FeatureVectorPCA

public FeatureVectorPCA(org.openimaj.math.matrix.algorithm.pca.PrincipalComponentAnalysis inner)

Construct with the given PrincipalComponentAnalysis object.

Parameters:

inner - PCA algorithm.

Method Detail

learnBasis

public void learnBasis(FeatureVector[] data)

Learn the PCA basis of the given feature vectors.

Parameters:

data - the feature vectors to apply PCA to.

learnBasis

public void learnBasis(Collection<? extends FeatureVector> data)

Learn the PCA basis of the given feature vectors.

Parameters:

data - the feature vectors to apply PCA to.

project

public DoubleFV project(FeatureVector vector)

Project a vector by the basis. The vector is normalised by subtracting the mean and then multiplied by the basis.

Parameters:

vector - the vector to project

Returns:

projected vector

learnBasis

public void learnBasis(double[][] data)

Overrides:

learnBasis in class org.openimaj.math.matrix.algorithm.pca.PrincipalComponentAnalysis

learnBasisNorm

protected void learnBasisNorm(Jama.Matrix norm)

Specified by:

learnBasisNorm in class org.openimaj.math.matrix.algorithm.pca.PrincipalComponentAnalysis

generate

public DoubleFV generate(DoubleFV scalings)

Generate a new "observation" as a linear combination of the principal components (PC): mean + PC * scaling. If the scaling vector is shorter than the number of components, it will be zero-padded. If it is longer, it will be truncated.

Parameters:

scalings - the weighting for each PC

Returns:

generated observation