creme Phishing

phishing.py

>>> from creme import compose
>>> from creme import datasets
>>> from creme import linear_model
>>> from creme import metrics
>>> from creme import preprocessing

>>> X_y = datasets.Phishing()

>>> model = compose.Pipeline(
...     preprocessing.StandardScaler(),
...     linear_model.LogisticRegression()
... )

>>> metric = metrics.Accuracy()

>>> for x, y in X_y:
...     y_pred = model.predict_one(x)      # make a prediction
...     metric = metric.update(y, y_pred)  # update the metric
...     model = model.fit_one(x, y)        # make the model learn

>>> metric
Accuracy: 89.28%

X_y is not instantiated anywhere! 😄

Good catch!