feature_engine

Big release

The new release adds all of the following:

• New transformer: datetime
• New transformer: selection by PSI
• New transformer: match variables
• New documentation layout: pydata template
• New documentation organisation: api and user guide
• Sponsorship
• Paper and citation

And a lot more expended functionality of current classes.

Fixes bug in one hot encoder

• Migrates repo to own organisation github.com/feature-engine
• adds new transformer (logcp)
• expands transformers with inverse_transform
• expands correlation selectors by allowing callables in the corr method

• enforce compatibility with Scikit-learn through check_estimator
• new transformer for cyclical features
• add categorical support for variables of type category
• add option to perform categorical transformation to numerical variables
• feature selectors support all cross-validation schemes

In this release we add:

• a new transformer for imputation
• a new transformer to create features
• new jupyter notebooks with demos
• linkd to Kaggle kernels featuring Feature-engine
• fix a number of typos in the documentation
  -fix few bugs in the select by target mean transformer.