# Package org.nd4j.linalg.lossfunctions.impl

• Class Summary
Class Description
LossBinaryXENT
Binary cross entropy loss function https://en.wikipedia.org/wiki/Cross_entropy#Cross-entropy_error_function_and_logistic_regression Labels are assumed to take values 0 or 1
LossCosineProximity
Created by susaneraly on 9/9/16.
LossFMeasure
F–measure loss function is a loss function design for training on imbalanced datasets.
LossHinge
Created by susaneraly on 8/15/16.
LossKLD
Kullback Leibler Divergence loss function
LossL1
L1 loss function: i.e., sum of absolute errors, L = sum_i abs(predicted_i - actual_i) See also `LossMAE` for a mathematically similar loss function (MAE has division by N, where N is output size)
LossL2
L2 loss function: i.e., sum of squared errors, L = sum_i (actual_i - predicted)^2 The L2 loss function is the square of the L2 norm of the difference between actual and predicted.
LossMAE
Mean absolute error loss function: L = 1/N sum_i abs(predicted_i - actual_i) See also `LossL1` for a mathematically similar loss function (LossL1 does not have division by N, where N is output size)
LossMAPE
Created by susaneraly on 8/15/16.
LossMCXENT
Multi-Class Cross Entropy loss function:
L = sum_i actual_i * log( predicted_i )
Note that labels are represented by a one-hot distribution
See `LossSparseMCXENT` for the equivalent but with labels as integers instead
LossMixtureDensity
This is a cost function associated with a mixture-density network.
LossMixtureDensity.Builder
LossMixtureDensity.MixtureDensityComponents
This class is a data holder for the mixture density components for convenient manipulation.
LossMSE
Mean Squared Error loss function: L = 1/N sum_i (actual_i - predicted)^2 See also `LossL2` for a mathematically similar loss function (LossL2 does not have division by N, where N is output size)
LossMSLE
Mean Squared Logarithmic Error loss function: L = 1/N sum_i (log(1+predicted_i) - log(1+actual_i))^2
LossMultiLabel
Multi-Label-Loss Function, maybe more commonly known as BPMLL This Loss function requires that the Labels are given as a multi-hot encoded vector.
LossNegativeLogLikelihood
Negative log likelihood loss function In practice, this is implemented as an alias for `LossMCXENT` due to the mathematical equivalence
LossPoisson
Created by susaneraly on 9/9/16.
LossSparseMCXENT
Sparse Multi-Class Cross Entropy loss function:
L = sum_i actual_i * log( predicted_i )
Note: this is the same loss function as `LossMCXENT`, the only difference being the format for the labels - this loss function uses integer indices (zero indexed) for the loss array, whereas LossMCXENT uses the equivalent one-hot representation
LossSquaredHinge
Created by susaneraly on 9/9/16.
LossWasserstein
Wasserstein loss function, which calculates the Wasserstein distance, also known as earthmover's distance.