Package org.deeplearning4j.nn.conf.ocnn

Class OCNNOutputLayer.Builder

java.lang.Object

org.deeplearning4j.nn.conf.layers.Layer.Builder<T>

org.deeplearning4j.nn.conf.layers.BaseLayer.Builder<T>

org.deeplearning4j.nn.conf.layers.FeedForwardLayer.Builder<T>

org.deeplearning4j.nn.conf.layers.BaseOutputLayer.Builder<OCNNOutputLayer.Builder>

org.deeplearning4j.nn.conf.ocnn.OCNNOutputLayer.Builder

Enclosing class:

OCNNOutputLayer

public static class OCNNOutputLayer.Builder
extends BaseOutputLayer.Builder<OCNNOutputLayer.Builder>

Field Summary

Fields

Modifier and Type	Field	Description
protected IActivation	activation	The activation function to use with ocnn
protected boolean	configureR	Whether to use the specified initialRValue or use the weight initialization with the neural network for the r value
protected int	hiddenLayerSize	The hidden layer size for the one class neural network.
protected double	initialRValue	The initial r value to use for ocnn for definition, see the paper, note this is only active when configureR is specified as true
protected double	nu	For nu definition see the paper
protected int	windowSize	The number of examples to use for computing the quantile for the r value update.

Fields inherited from class org.deeplearning4j.nn.conf.layers.BaseOutputLayer.Builder

lossFn

Fields inherited from class org.deeplearning4j.nn.conf.layers.FeedForwardLayer.Builder

nIn, nOut

Fields inherited from class org.deeplearning4j.nn.conf.layers.BaseLayer.Builder

activationFn, biasInit, biasUpdater, gainInit, gradientNormalization, gradientNormalizationThreshold, iupdater, regularization, regularizationBias, weightInitFn, weightNoise

Fields inherited from class org.deeplearning4j.nn.conf.layers.Layer.Builder

allParamConstraints, biasConstraints, iDropout, layerName, weightConstraints

Constructor Summary

Constructors

Constructor	Description
Builder()

Method Summary

Modifier and Type	Method	Description
OCNNOutputLayer.Builder	activation(IActivation activation)	The activation function to use with ocnn
OCNNOutputLayer	build()
OCNNOutputLayer.Builder	configureR(boolean configureR)	Whether to use the specified initialRValue or use the weight initialization with the neural network for the r value
OCNNOutputLayer.Builder	hiddenLayerSize(int hiddenLayerSize)	The hidden layer size for the one class neural network.
OCNNOutputLayer.Builder	initialRValue(double initialRValue)	The initial r value to use for ocnn for definition, see the paper, note this is only active when configureR is specified as true
OCNNOutputLayer.Builder	nOut(int nOut)	Number of outputs - used to set the layer size (number of units/nodes for the current layer).
OCNNOutputLayer.Builder	nu(double nu)	For nu definition see the paper
void	setNOut(long nOut)
OCNNOutputLayer.Builder	windowSize(int windowSize)	The number of examples to use for computing the quantile for the r value update.

Methods inherited from class org.deeplearning4j.nn.conf.layers.BaseOutputLayer.Builder

hasBias, lossFunction, lossFunction

Methods inherited from class org.deeplearning4j.nn.conf.layers.FeedForwardLayer.Builder

nIn, nIn, nOut, units

Methods inherited from class org.deeplearning4j.nn.conf.layers.BaseLayer.Builder

activation, biasInit, biasUpdater, dist, gainInit, gradientNormalization, gradientNormalizationThreshold, l1, l1Bias, l2, l2Bias, regularization, regularizationBias, updater, updater, weightDecay, weightDecay, weightDecayBias, weightDecayBias, weightInit, weightInit, weightInit, weightNoise

Methods inherited from class org.deeplearning4j.nn.conf.layers.Layer.Builder

constrainAllParameters, constrainBias, constrainWeights, dropOut, dropOut, name

Methods inherited from class java.lang.Object

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

Field Detail

hiddenLayerSize

protected int hiddenLayerSize

The hidden layer size for the one class neural network. Note this would be nOut on a dense layer. NOut in this neural net is always set to 1 though.

nu

protected double nu

For nu definition see the paper

windowSize

protected int windowSize

The number of examples to use for computing the quantile for the r value update. This value should generally be the same as the number of examples in the dataset

activation

protected IActivation activation

The activation function to use with ocnn

initialRValue

protected double initialRValue

The initial r value to use for ocnn for definition, see the paper, note this is only active when configureR is specified as true

configureR

protected boolean configureR

Whether to use the specified initialRValue or use the weight initialization with the neural network for the r value

Constructor Detail

Builder

public Builder()

Method Detail

configureR

public OCNNOutputLayer.Builder configureR(boolean configureR)

Whether to use the specified initialRValue or use the weight initialization with the neural network for the r value

Parameters:

configureR - true if we should use the initial initialRValue

initialRValue

public OCNNOutputLayer.Builder initialRValue(double initialRValue)

The initial r value to use for ocnn for definition, see the paper, note this is only active when configureR is specified as true

Parameters:

initialRValue - the int

windowSize

public OCNNOutputLayer.Builder windowSize(int windowSize)

The number of examples to use for computing the quantile for the r value update. This value should generally be the same as the number of examples in the dataset

Parameters:

windowSize - the number of examples to use for computing the quantile of the dataset for the r value update

nu

public OCNNOutputLayer.Builder nu(double nu)

For nu definition see the paper

Parameters:

nu - the nu for ocnn

activation

public OCNNOutputLayer.Builder activation(IActivation activation)

The activation function to use with ocnn

Overrides:

activation in class BaseLayer.Builder<OCNNOutputLayer.Builder>

Parameters:

activation - the activation function to sue

hiddenLayerSize

public OCNNOutputLayer.Builder hiddenLayerSize(int hiddenLayerSize)

The hidden layer size for the one class neural network. Note this would be nOut on a dense layer. NOut in this neural net is always set to 1 though.

Parameters:

hiddenLayerSize - the hidden layer size to use with ocnn

nOut

public OCNNOutputLayer.Builder nOut(int nOut)

Description copied from class: FeedForwardLayer.Builder

Number of outputs - used to set the layer size (number of units/nodes for the current layer). Note that this is equivalent to FeedForwardLayer.Builder.units(int)

Overrides:

nOut in class FeedForwardLayer.Builder<OCNNOutputLayer.Builder>

Parameters:

nOut - Number of outputs / layer size

setNOut

public void setNOut(long nOut)

build

public OCNNOutputLayer build()

Specified by:

build in class Layer.Builder<OCNNOutputLayer.Builder>