boosters.Dataset#

class boosters.Dataset[source]#

Bases: Dataset

Dataset holding features, labels, and optional metadata.

was_converted: bool#
static __new__(cls, features, labels=None, weights=None, feature_names=None, categorical_features=None)[source]#

Create a Dataset from numpy arrays or a pandas DataFrame.

Return type:

Dataset

Parameters:

  • features (object)

  • labels (LabelsInput)

  • weights (WeightsInput)

  • feature_names (Sequence[str] | None)

  • categorical_features (Sequence[int] | None)

categorical_features#

Indices of categorical features.

feature_names#

Feature names if provided.

has_labels#

Whether labels are present.

has_weights#

Whether weights are present.

n_features#

Number of features in the dataset.

n_samples#

Number of samples in the dataset.

shape#

Shape of the features array as (n_samples, n_features).

__init__(features, labels=None, weights=None, feature_names=None, categorical_features=None)[source]#

No-op: all initialization is performed in __new__.

Parameters:

  • features (object)

  • labels (LabelsInput)

  • weights (WeightsInput)

  • feature_names (Sequence[str] | None)

  • categorical_features (Sequence[int] | None)

Return type:

None

classmethod builder()[source]#

Create a builder for step-by-step dataset construction.

Return type:

DatasetBuilder