Defines a specific data analysis model, including features, estimator and target metric. Can be stored (pickled) and retrieved.
Args
target: Feature or basestring specifying the target (“y”) variable of the analysis.
features: an iterable of Features <Feature> to be used by the estimator in the analysis.
model: an estimator instance compatible with sklearn estimator conventions. (has fit(x, y) and predict(y) methods).
metrics: an iterable of evaluation `Metric`s used to score predictions.
reporters: an iterable of Reporter objects
prediction: a Feature transformation of the special predictions_name column used to post-process predictions prior to metric scoring.
predictions_name: a unique string used as a column identifier for model predictions. Must be unique among all feature names: eg ‘$logreg_predictions$’
Provides an iterator over passed in configuration values, allowing for easy exploration of models.
Args
base_config: The base Configuration to augment
kwargs: Can be any keyword accepted by Configuration. Values should be iterables.