FTRL

Online-style follow-the-regularized-leader classifier

Code 
from _api_doc_utils import *

1 Where it fits

Group: Regression

FTRL is an online-style regularized classifier. The update combines adaptive learning rates with L1/L2 regularization, making it useful for sparse or streaming-style binary prediction problems.

The exposed class is deliberately simple: fit once on a design matrix and binary labels, then predict scores for new rows.

2 Python API

Constructor: cm.FTRL

Call FTRL(alpha, beta, l1_ratio, l2_ratio), then fit(x, y_int32). predict(x) returns scores in [0, 1]; summary() returns coefficient estimates and marks analytic inference unavailable for the penalized online estimator.

Code 
print(inspect.signature(cm.FTRL))
Code 
cls = cm.FTRL
display(HTML(html_table(["Public method"], public_methods(cls))))

3 Minimal example

Code 
rng = np.random.default_rng(8)
x = rng.normal(size=(220, 4))
p = 1 / (1 + np.exp(-(x @ np.array([0.7, -0.4, 0.2, 0.3]))))
y = rng.binomial(1, p, size=220).astype(np.int32)
model = cm.FTRL()
model.fit(x, y)
print(model.summary()['coef'])
print(model.predict(x[:5]))

4 summary() contract

The table below is generated by fitting the live class in this repository and then inspecting summary(). Shapes are shown because most values are plain NumPy arrays or scalars.

Code 
rng = np.random.default_rng(108)
x = rng.normal(size=(100, 4))
y = rng.binomial(1, 0.5, size=100).astype(np.int32)
model = cm.FTRL()
model.fit(x, y)
summary = model.summary()
display(HTML(html_table(["summary() key", "shape"], summary_shape_rows(summary))))