#6372 adds get_feature_names to PolynomialFeatures. It accepts a list of names of input_features (or substitutes with defaults) and constructs feature name strings that are human-readable and informative. Similar support should be available for other transformers, including feature selectors, feature agglomeration, FunctionTransformer, and perhaps even PCA (giving the top contributors to each component). FeatureUnion should be modified to handle the case where an argument is supplied. A proposal for support in Pipeline is given in #6424.

Modelled on #6372, each enhancement can be contributed as a separate PR. Note that default names for features are [x0, x1, ...]

• PolynomialFeatures @amueller [MRG+2] add get_feature_names to PolynomialFeatures #6372
• feature selection and randomized L1 @yenchenlin1994
• feature agglomeration @yenchenlin1994
• FunctionTransformer @nelson-liu [MRG] ENH Add get_feature_names for various transformers #6431
• scalers, normalizers and imputers: should be trivial like FunctionTransformer [MRG] ENH Add get_feature_names for various transformers #6431
• Binarizer [MRG] ENH Add get_feature_names for various transformers #6431
• OneHotEncoder [MRG] ENH Add get_feature_names for OneHotEncoder #6441
• FeatureUnion @yenchenlin1994
• PCA?
• SparseRandomProjection?
• GaussianRandomProjection??
• other transformers?