org.nd4j.linalg.lossfunctions

Class LossFunctions

java.lang.Object

org.nd4j.linalg.lossfunctions.LossFunctions

public class LossFunctions
extends java.lang.Object

Central class for loss functions

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	LossFunctions.LossFunction MSE: Mean Squared Error: Linear Regression EXPLL: Exponential log likelihood: Poisson Regression XENT: Cross Entropy: Binary Classification SOFTMAX: Softmax Regression RMSE_XENT: RMSE Cross Entropy

Constructor Summary

Constructors

Constructor and Description
LossFunctions()

Method Summary

Modifier and Type	Method and Description
static double	reconEntropy(INDArray input, INDArray hBias, INDArray vBias, INDArray W, ActivationFunction activationFunction) Reconstruction entropy for Denoising AutoEncoders and RBMs
static double	score(INDArray labels, LossFunctions.LossFunction lossFunction, INDArray output, double l2, boolean useRegularization) Generic scoring function

Methods inherited from class java.lang.Object

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

Constructor Detail

LossFunctions

public LossFunctions()

Method Detail

score

public static double score(INDArray labels,
                           LossFunctions.LossFunction lossFunction,
                           INDArray output,
                           double l2,
                           boolean useRegularization)

Generic scoring function

Parameters:

labels - the labels to score

lossFunction - the loss function to use

output - the output function

l2 - the l2 coefficient

useRegularization - whether to use regularization

Returns:

the score for the given parameters

reconEntropy

public static double reconEntropy(INDArray input,
                                  INDArray hBias,
                                  INDArray vBias,
                                  INDArray W,
                                  ActivationFunction activationFunction)

Reconstruction entropy for Denoising AutoEncoders and RBMs

Parameters:

input - the input ndarray

hBias - the hidden bias of the neural network

vBias - the visible bias of the neural network

W - the weight matrix of the neural network

Returns:

the reconstruction cross entropy for the given parameters