Login
Back to bricks list
Introduction Version

SGD classifier trainer

TASK
Typing name :  TASK.gws_gaia.SGDClassifierTrainer Brick :  gws_gaia

Train a stochastic gradient descent (SGD) linear classifier

Trainer of a linear classifier with stochastic gradient descent (SGD). Fit a SGD linear classifier with a training table.

See https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.SGDClassifier.html for more details.

Input

Table
The input table

Output

result
The output result

Configuration

training_design

Define the training design, i.e. the target Y to use for the model.

Type : ListMaximum occurrences number : -1

target_name

The name of the 'columns' or 'row_tag keys' to use as targets or labels.

Type : string

target_origin

Optional

The origin of the target. Notice: Targets comming from a 'row_tag' are always considered as categorical.

Type : stringAllowed values : column  row_tag  Default value : column

target_type

Optional

The type of the target (categorical or numerical). Set 'auto' to infer the correct type. Notice: targets comming from row_tags are allways considered as categorical

Type : stringAllowed values : auto  categorical  numerical  Default value : auto

loss

Optional

Type : stringAllowed values : hinge  log_loss  log  modified_huber  squared_hinge  perceptron  squared_error  huber  epsilon_insensitive  squared_epsilon_insensitive  Default value : hinge

alpha

Optional

Type : floatDefault value : 0.0001

max_iter

Optional

Type : intDefault value : 1000