feature#

Feature selection.

Feature selection identifies the subset of features that are most relevant to the learning task. In data stream learning, feature relevance can change over time, so importance must be estimated and updated incrementally.

Classes#

FeatureImportanceClassifier

Base class for classifiers that expose feature-importance estimates.

MOAFeatureImportanceClassifier

MOA-backed feature-importance classifier.

Functions#

plot_feature_importance

Plot feature importances as a bar chart.

plot_windowed_feature_importance

Plot windowed feature importances over time.
capymoa.feature.plot_feature_importance(
importances: Sequence[float],
feature_names: Sequence[str] | str | None = None,
*,
top_k: int | None = None,
ax=None,
title: str = 'Feature importances',
)[source]#

Plot feature importances as a bar chart.

capymoa.feature.plot_windowed_feature_importance(
windowed_importances: list[dict],
feature_names: Sequence[str] | str | None = None,
*,
top_k: int | None = None,
ax=None,
title: str = 'Windowed feature importances',
)[source]#

Plot windowed feature importances over time.