datatable.models.LinearModel

This class implements the Linear model with the stochastic gradient descent learning. It supports linear regression, as well as binomial and multinomial classification. Both .fit() and .predict() methods are fully parallel.

Construction

LinearModel()

Construct a LinearModel object.

Methods

fit()

Train model on the input samples and targets.

is_fitted()

Report model status.

predict()

Predict for the input samples.

reset()

Reset the model.

Properties

eta0

Initial learning rate.

eta_decay

Decay for the "time-based" and "step-based" learning rate schedules.

eta_drop_rate

Drop rate for the "step-based" learning rate schedule.

eta_schedule

Learning rate schedule.

double_precision

An option to control precision of the internal computations.

labels

Classification labels.

lambda1

L1 regularization parameter.

lambda2

L2 regularization parameter.

model

Model coefficients.

model_type

Model type to be built.

negative_class

An option to indicate if the “negative” class should be a created for multinomial classification.

nepochs

Number of training epochs.

params

All the input model parameters as a named tuple.

seed

Seed for the quasi-random data shuffling.