PandasDataset

class lightautoml.dataset.np_pd_dataset.PandasDataset(data=None, roles=None, task=None, **kwargs)

Bases: lightautoml.dataset.base.LAMLDataset

Dataset that contains pd.DataFrame features and pd.Series targets.

property features

Get list of features.

Return type

List[str]

Returns

list of features.

__init__(data=None, roles=None, task=None, **kwargs)

Create dataset from pd.DataFrame and pd.Series.

Parameters

• data (Optional[DataFrame]) – Table with features.

• features – features names.

• roles (Optional[Dict[str, ColumnRole]]) – Roles specifier.

• task (Optional[Task]) – Task specifier.

• **kwargs – Series, array like attrs target, group etc…

set_data(data, features, roles)

Inplace set data, features, roles for empty dataset.

Parameters

• data (DataFrame) – Table with features.

• features (None) – None, just for same interface.

• roles (Dict[str, ColumnRole]) – Dict with roles.

to_numpy()

Convert to class:NumpyDataset.

Returns

NumpyDataset format.

Return type

Same dataset in class

to_pandas()

Empty method, return the same object.

Return type

PandasDataset

Returns

Self.

static from_dataset(dataset)

Convert random dataset to pandas dataset.

Return type

PandasDataset

Returns

Converted to pandas dataset.

nan_rate()

Counts overall number of nans in dataset.

Returns

Number of nans.