Beta Version 1.13.0

Added

• StateActionRewardStateActionNeuralNetwork

Fixes

• None

Changes

• Non-binary classification models now have the option to return the original output values from prediction in the form of matrix. You can activate this by adding “true” after featureMatrix parameters in predict() function.

• Neural network now accepts labelMatrix in train() function. Just replace the labelVector with labelMatrix.

• Naive Bayes will now return highest probability as opposed to probability vector.