Tsne (bh-tsne 0.0.1 API)

java.lang.Object
- de.javagl.tsne.Tsne

```
public final class Tsne
extends java.lang.Object
```
A class offering the interface to the underlying t-SNE implementation.

Constructor Summary

Constructors
Constructor and Description

Tsne()
Creates a new instance.

Constructors
Constructor and Description
`Tsne()` Creates a new instance.

Method Summary

All Methods Instance Methods Concrete Methods
Modifier and Type	Method and Description
`int`	`getInitialDims()` Returns the initial dimensions.
`int`	`getMaxIterations()` Returns the maximum number of iterations that will be performed.
`java.util.function.Consumer<? super java.lang.String>`	`getMessageConsumer()` Returns the consumer for the messages
`int`	`getOutputDims()` Returns the number of output dimensions
`double`	`getPerplexity()` Returns the perplexity.
`java.util.function.DoubleConsumer`	`getProgressConsumer()` Returns the progress consumer
`long`	`getRandomSeed()` Returns the random seed
`double`	`getTheta()` Returns the theta parameter.
`double[][]`	`run(double[][] input)` Performs the t-SNE on the given input data, and returns the result.
`void`	`setInitialDims(int initialDims)` Set the number of initial dimensions.
`void`	`setMaxIterations(int maxIterations)` Set the maximum number of iterations that will be performed.
`void`	`setMessageConsumer(java.util.function.Consumer<? super java.lang.String> messageConsumer)` Set an optional consumer for the messages that are emitted by the underlying implementation.
`void`	`setOutputDims(int outputDims)` Set the output dimensions.
`void`	`setPerplexity(double perplexity)` Set the perplexity.
`void`	`setProgressConsumer(java.util.function.DoubleConsumer progressConsumer)` Set an optional consumer for the progress of the iterations.
`void`	`setRandomSeed(long randomSeed)` Set the random seed.
`void`	`setTheta(double theta)` Set this theta parameter.

Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

- Constructor Detail
  - Tsne
```
public Tsne()
```
    Creates a new instance.
    
    The output dimensions will be 2.
    The maximum iterations will be 1000.
    The perplexity will be 20.0.
    The theta will be 0.5.
    The initial dimensions will be 55.
    The random seed will be 0.
    
    Messages will be written to the standard output. To disable the output, call setMessageConsumer(null).
- Method Detail
  - setOutputDims
```
public void setOutputDims(int outputDims)
```
    Set the output dimensions. This should (or will) be 2 in most cases.
    
    Parameters:
    
    outputDims - The output dimensions
    
    Throws:
    
    java.lang.IllegalArgumentException - If the given number is not positive
  - getOutputDims
```
public int getOutputDims()
```
    Returns the number of output dimensions
    
    Returns:
    
    The number of output dimensions
  - setMaxIterations
```
public void setMaxIterations(int maxIterations)
```
    Set the maximum number of iterations that will be performed.
    
    Note that due to some learning parameters that are used in the underlying implementation (and that are not accessible from the outside), the number should be ~"considerably larger" than 250.
    
    Parameters:
    
    maxIterations - The maximum number of iterations
    
    Throws:
    
    java.lang.IllegalArgumentException - If the given number is not positive
  - getMaxIterations
```
public int getMaxIterations()
```
    Returns the maximum number of iterations that will be performed. See setMaxIterations(int) for details.
    
    Returns:
    
    The maximum number of iterations.
  - setInitialDims
```
public void setInitialDims(int initialDims)
```
    Set the number of initial dimensions.
    
    This is the number of dimensions that the input data should be reduced to (using a PCA). If this value is not positive, or if it is larger than the dimensionality of the actual input data, then no PCA will be performed and the dimensionality of the input data will be used.
    
    Parameters:
    
    initialDims - The initial dimensions
  - getInitialDims
```
public int getInitialDims()
```
    Returns the initial dimensions. See setInitialDims(int) for details.
    
    Returns:
    
    The initial dimensions.
  - setPerplexity
```
public void setPerplexity(double perplexity)
```
    Set the perplexity.
    
    According to the t-SNE FAQ, this can be seen as a measure for the number of effective nearest neighbors. It should be larger for larger/denser data sets. The typical value range is between 5.0 and 50.0.
    
    Note that numberOfInputPoints - 1 > 3 * perplexity must hold, otherwise, the underlying implementation will throw an exception.
    
    Parameters:
    
    perplexity - The perplexity
  - getPerplexity
```
public double getPerplexity()
```
    Returns the perplexity. See setPerplexity(double) for details.
    
    Returns:
    
    The perplexity.
  - setTheta
```
public void setTheta(double theta)
```
    Set this theta parameter. Don't do this if you don't know what it does.
    
    Parameters:
    
    theta - The theta parameter.
    
    Throws:
    
    java.lang.IllegalArgumentException - If the given number is not positive
  - getTheta
```
public double getTheta()
```
    Returns the theta parameter. See setTheta(double) for details.
    
    Returns:
    
    The theta parameter
  - setRandomSeed
```
public void setRandomSeed(long randomSeed)
```
    Set the random seed. By default, this will be 0, causing the same results to be generated each time. In order to generate different results, the random seed may be set to different values.
    
    Parameters:
    
    randomSeed - The random seed
  - getRandomSeed
```
public long getRandomSeed()
```
    Returns the random seed
    
    Returns:
    
    The random seed
  - setProgressConsumer
```
public void setProgressConsumer(java.util.function.DoubleConsumer progressConsumer)
```
    Set an optional consumer for the progress of the iterations. This will receive a value in [0.0, 1.0] indicating the progress of the iterations.
    
    Parameters:
    
    progressConsumer - The progress consumer
  - getProgressConsumer
```
public java.util.function.DoubleConsumer getProgressConsumer()
```
    Returns the progress consumer
    
    Returns:
    
    The progress consumer
  - setMessageConsumer
```
public void setMessageConsumer(java.util.function.Consumer<? super java.lang.String> messageConsumer)
```
    Set an optional consumer for the messages that are emitted by the underlying implementation.
    
    Parameters:
    
    messageConsumer - The message consumer
  - getMessageConsumer
```
public java.util.function.Consumer<? super java.lang.String> getMessageConsumer()
```
    Returns the consumer for the messages
    
    Returns:
    
    The message consumer
  - run
```
public double[][] run(double[][] input)
```
    Performs the t-SNE on the given input data, and returns the result.
    
    Parameters:
    
    input - The input data
    
    Returns:
    
    The result, or null if the calling thread was interrupted

Class Tsne

Constructor Summary

Method Summary

Methods inherited from class java.lang.Object

Constructor Detail

Tsne

Method Detail

setOutputDims

getOutputDims

setMaxIterations

getMaxIterations

setInitialDims

getInitialDims

setPerplexity

getPerplexity

setTheta

getTheta

setRandomSeed

getRandomSeed

setProgressConsumer

getProgressConsumer

setMessageConsumer

getMessageConsumer

run