org.apache.spark.streamdm.classifiers
Implementation of the squared loss function.
The L1Regularizer gradient return the sign as the weight for the regularization.
The L2Regularizer gradient returns the weight as the gradient for the regularization.
A Model trait defines the needed operations on any learning Model.
Implementation of the logistic loss function.
A Loss trait defines the operation needed to compute the loss function, the prediction function, and the gradient for use in a LinearModel.
Implementation of the perceptron loss function.
A regularizer trait defines the gradient operation for computing regularized models.
Implementation of the squared loss function.
The ZeroRegularizer gradient simply returns 0 as the gradient.