BestOfTransformers
- class lightautoml.transformers.base.BestOfTransformers(transformer_list, criterion)[source]
Bases:
lightautoml.transformers.base.LAMLTransformerApply multiple transformers and select best.
- __init__(transformer_list, criterion)[source]
Create selector from candidate list and selection criterion.
- Parameters
transformer_list (
Sequence[LAMLTransformer]) – Sequence of transformers.criterion (
Callable) – Score function (greater is better).
- fit(dataset)[source]
- Empty method - raise error.
This transformer supports only
fit_transform.
- Parameters
dataset (
LAMLDataset) – LAMLDataset to fit on.- Raises
NotImplementedError – Always.
- fit_transform(dataset)[source]
Fit transform all transformers and then select best.
- Parameters
dataset (
LAMLDataset) – Dataset to fit and then transform.- Return type
- Returns
Dataset with new features.
- transform(dataset)[source]
Make transform by the best selected transformer.
- Parameters
dataset (
LAMLDataset) – Dataset with input features.- Return type
- Returns
Dataset with new features.