Exp Class

Some of the trainers accept a loss parameter that will be used for training. It is also known as loss function, objective function, or optimization score function.

Inheritance

builtins.object

Exp

Constructor

Exp(beta=1.0)

Parameters

Name	Description
beta	Dilation

Examples

   ###############################################################################
   # Exponential Loss
   from nimbusml.linear_model import SgdBinaryClassifier
   from nimbusml.loss import Exp

   # specify loss function using string keyword
   trainer1 = SgdBinaryClassifier(loss='exp')

   # can also use the loss class instead of string.

   trainer1 = SgdBinaryClassifier(loss=Exp())  # equivalent to loss='exp'
   trainer2 = SgdBinaryClassifier(loss=Exp(beta=0.4))

Remarks

Losses can be specified either as a string or a loss object. When loss is specified as one of these strings, the default values are used for the loss parameters. To change the default parameters, a loss object should be used, as seen in examples below.

Each trainer supports only a subset of the losses mentioned above. To get the supported losses and the default loss, please refer to the documentation page for the specific trainer.

The Exponential loss for classification. Compared to Hinge, it penalizes more on the wrong prediction and has larger gradients. Its string name is 'exp'.

It can be used for AveragedPerceptronBinaryClassifier, SgdBinaryClassifier.