org.deeplearning4j.plot

Class Tsne

java.lang.Object

org.deeplearning4j.plot.Tsne

All Implemented Interfaces:

Serializable

Direct Known Subclasses:

BarnesHutTsne

public class Tsne
extends Object
implements Serializable

Tsne calculation

Author:

Adam Gibson

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	Tsne.Builder

Field Summary

Fields

Modifier and Type	Field and Description
protected org.nd4j.linalg.learning.AdaGrad	adaGrad
protected double	finalMomentum
protected org.nd4j.linalg.api.ndarray.INDArray	gains
protected double	initialMomentum
protected IterationListener	iterationListener
protected double	learningRate
protected static org.slf4j.Logger	log
protected int	maxIter
protected double	minGain
protected double	momentum
protected boolean	normalize
protected double	perplexity
protected static org.springframework.core.io.ClassPathResource	r
protected double	realMin
protected int	stopLyingIteration
protected int	switchMomentumIteration
protected double	tolerance
protected boolean	useAdaGrad
protected boolean	usePca
protected org.nd4j.linalg.api.ndarray.INDArray	y
protected org.nd4j.linalg.api.ndarray.INDArray	yIncs

Constructor Summary

Constructors

Constructor and Description
Tsne()
Tsne(int maxIter, double realMin, double initialMomentum, double finalMomentum, double momentum, int switchMomentumIteration, boolean normalize, boolean usePca, int stopLyingIteration, double tolerance, double learningRate, boolean useAdaGrad, double perplexity, double minGain)

Method Summary

Modifier and Type	Method and Description
org.nd4j.linalg.api.ndarray.INDArray	calculate(org.nd4j.linalg.api.ndarray.INDArray X, int nDims, double perplexity)
org.nd4j.linalg.api.ndarray.INDArray	computeGaussianPerplexity(org.nd4j.linalg.api.ndarray.INDArray d, double u) Convert data to probability co-occurrences (aka calculating the kernel)
IterationListener	getIterationListener()
org.nd4j.linalg.api.ndarray.INDArray	getY()
org.nd4j.linalg.api.ndarray.INDArray	getYGradient(int n, org.nd4j.linalg.api.ndarray.INDArray PQ, org.nd4j.linalg.api.ndarray.INDArray qu)
protected Pair<Double,org.nd4j.linalg.api.ndarray.INDArray>	gradient(org.nd4j.linalg.api.ndarray.INDArray p)
Pair<org.nd4j.linalg.api.ndarray.INDArray,org.nd4j.linalg.api.ndarray.INDArray>	hBeta(org.nd4j.linalg.api.ndarray.INDArray d, double beta) Computes a gaussian kernel given a vector of squared distance distances
void	plot(org.nd4j.linalg.api.ndarray.INDArray matrix, int nDims, List<String> labels) Plot tsne (write the coordinates file)
void	plot(org.nd4j.linalg.api.ndarray.INDArray matrix, int nDims, List<String> labels, String path) Plot tsne
void	setIterationListener(IterationListener iterationListener)
void	setY(org.nd4j.linalg.api.ndarray.INDArray y)
void	step(org.nd4j.linalg.api.ndarray.INDArray p, int i) An individual iteration

Methods inherited from class java.lang.Object

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

Field Detail

maxIter

protected int maxIter

realMin

protected double realMin

initialMomentum

protected double initialMomentum

finalMomentum

protected double finalMomentum

minGain

protected double minGain

momentum

protected double momentum

switchMomentumIteration

protected int switchMomentumIteration

normalize

protected boolean normalize

usePca

protected boolean usePca

stopLyingIteration

protected int stopLyingIteration

tolerance

protected double tolerance

learningRate

protected double learningRate

adaGrad

protected org.nd4j.linalg.learning.AdaGrad adaGrad

useAdaGrad

protected boolean useAdaGrad

perplexity

protected double perplexity

gains

protected org.nd4j.linalg.api.ndarray.INDArray gains

yIncs

protected org.nd4j.linalg.api.ndarray.INDArray yIncs

y

protected org.nd4j.linalg.api.ndarray.INDArray y

iterationListener

protected transient IterationListener iterationListener

r

protected static org.springframework.core.io.ClassPathResource r

log

protected static final org.slf4j.Logger log

Constructor Detail

Tsne

public Tsne()

Tsne

public Tsne(int maxIter,
            double realMin,
            double initialMomentum,
            double finalMomentum,
            double momentum,
            int switchMomentumIteration,
            boolean normalize,
            boolean usePca,
            int stopLyingIteration,
            double tolerance,
            double learningRate,
            boolean useAdaGrad,
            double perplexity,
            double minGain)

Method Detail

hBeta

public Pair<org.nd4j.linalg.api.ndarray.INDArray,org.nd4j.linalg.api.ndarray.INDArray> hBeta(org.nd4j.linalg.api.ndarray.INDArray d,
                                                                                             double beta)

Computes a gaussian kernel given a vector of squared distance distances

Parameters:

d - the data

beta -

Returns:

computeGaussianPerplexity

public org.nd4j.linalg.api.ndarray.INDArray computeGaussianPerplexity(org.nd4j.linalg.api.ndarray.INDArray d,
                                                                      double u)

Convert data to probability co-occurrences (aka calculating the kernel)

Parameters:

d - the data to convert

u - the perplexity of the model

Returns:

the probabilities of co-occurrence

calculate

public org.nd4j.linalg.api.ndarray.INDArray calculate(org.nd4j.linalg.api.ndarray.INDArray X,
                                                      int nDims,
                                                      double perplexity)

Parameters:

X -

nDims -

perplexity -

gradient

protected Pair<Double,org.nd4j.linalg.api.ndarray.INDArray> gradient(org.nd4j.linalg.api.ndarray.INDArray p)

getYGradient

public org.nd4j.linalg.api.ndarray.INDArray getYGradient(int n,
                                                         org.nd4j.linalg.api.ndarray.INDArray PQ,
                                                         org.nd4j.linalg.api.ndarray.INDArray qu)

step

public void step(org.nd4j.linalg.api.ndarray.INDArray p,
                 int i)

An individual iteration

Parameters:

p - the probabilities that certain points are near each other

i - the iteration (primarily for debugging purposes)

plot

public void plot(org.nd4j.linalg.api.ndarray.INDArray matrix,
                 int nDims,
                 List<String> labels)
          throws IOException

Plot tsne (write the coordinates file)

Parameters:

matrix - the matrix to plot

nDims - the number of dimensions

labels -

Throws:

IOException

plot

public void plot(org.nd4j.linalg.api.ndarray.INDArray matrix,
                 int nDims,
                 List<String> labels,
                 String path)
          throws IOException

Plot tsne

Parameters:

matrix - the matrix to plot

nDims - the number

labels -

path - the path to write

Throws:

IOException

getY

public org.nd4j.linalg.api.ndarray.INDArray getY()

setY

public void setY(org.nd4j.linalg.api.ndarray.INDArray y)

getIterationListener

public IterationListener getIterationListener()

setIterationListener

public void setIterationListener(IterationListener iterationListener)