pyreal.RealApp.from_sklearn#

static RealApp.from_sklearn(pipeline=None, model=None, transformers=None, X_train=None, y_train=None, refit_model=True, verbose=0, **kwargs)[source]#

Create a RealApp from a sklearn pipeline or model and transformers. Must provide one of:

  • just pipeline

  • just model

  • model and transformers

Parameters:

  • pipeline (sklearn.pipeline.Pipeline) – Sklearn pipeline to convert. The final step of the pipeline must be a model.

  • model (sklearn model) – Sklearn model to use. Ignored if pipeline is not None

  • transformers (list of sklearn transformers) – List of sklearn transformers to use. Ignored if pipeline is not None

  • X_train (DataFrame) – Training data to fit transformers and explanations to. May be required if transformers are not fitted or must be recreated. If not provided, must be provided when preparing and using realapp explainers.

  • y_train (DataFrame or Series) – Training targets to fit transformers and explanations to. If not provided, must be provided when preparing and using realapp explainers.

  • refit_model (bool) – If True, refit the model using the new Pyreal transformers. This may be necessary as sklearn and Pyreal transformers may result in an unaligned column order. Requires X_train and y_train to be provided.

  • verbose (int) – Verbosity level. If 0, no output. If 1, detailed output

  • **kwargs – Additional arguments to pass to RealApp constructor.

Returns:

RealApp

Newly created RealApp object