README.md

Jupyter notebooks with Demos of Feature-engine's functionality

Feature-engine is a Python library with multiple transformers to engineer and select features for use in machine learning models. Feature-engine's transformers follow scikit-learn's functionality with fit() and transform() methods to first learn the transforming parameters from data and then transform the data.

In this repo, you will find a lot of examples on how to use Feature-engine's transformers on various datasets. The notebooks are sorted in the following folders and include examples for the following transformers:

creation

• MathFeatures
• RelativeFeatures
• CyclicalFeatures

discretisation

• EqualFrequencyDiscretiser
• EqualFrequencyDiscretiser plus WoEEncoder
• EqualWidthDiscretiser
• EqualWidthDiscretiser plus OrdinalEncoder
• DecisionTreeDiscretiser
• ArbitraryDiscreriser
• ArbitraryDiscreriser plus MeanEncoder

encoding

• OneHotEncoder
• OrdinalEncoder
• CountFrequencyEncoder
• MeanEncoder
• WoEEncoder
• PRatioEncoder
• RareLabelEncoder
• DecisionTreeEncoder

imputation

• MeanMedianImputer
• RandomSampleImputer
• EndTailImputer
• AddMissingIndicator
• CategoricalImputer
• ArbitraryNumberImputer
• DropMissingData -- notebook wanted, please contribute

outliers

• Winsorizer
• ArbitraryOutlierCapper
• OutlierTrimmer

pipelines

• create new features - wine data
• regression pipeline - house prices data
• more notebooks wanted, please constribute

transformation

• LogTransformer
• LogCpTransformer
• ReciprocalTransformer
• PowerTransformer
• BoxCoxTransformer
• YeoJohnsonTransformer

wrappers

• SklearnTransformerWrapper plus Scikit-learn's OneHotEncoder
• SklearnTransformerWrapper plus Scikit-learn's feature selection classes
• SklearnTransformerWrapper plus Scikit-learn's KBinsDiscretizer
• SklearnTransformerWrapper plus Scikit-learn's Scalers
• SklearnTransformerWrapper plus Scikit-learn's SimpleImputer

selection

• notebooks wanted, please contribute

Contributing

We welcome notebooks from users of the package. If you want to create one of the missing notebooks, or want to add a notebook of your own, provided that the data set is free to share, make a pull request with the code.

How to contribute:

Local Setup Steps

• Fork the repo
• Clone your fork into your local computer: git clone https://github.com/<YOURUSERNAME>/feature-engine-examples.git
• cd into the repo cd feature-engine-examples
• If you haven't done so yet, install feature-engine pip install feature_engine
• Create a feature branch with a meaningful name for your notebook: git checkout -b mynotebookbranch
• Develop your notebook
• Add the changes to your copy of the fork: git add ., git commit -m "a meaningful commit message", git pull origin mynotebookbranch:mynotebookbranch
• Go to your fork on Github and make a PR to this repo
• Done

Thank you!!

Feature-engine features in the following resources:

• Feature Engineering for Machine Learning, Online Course

• Feature Selection for Machine Learning, Online Course

• Python Feature Engineering Cookbook

Blogs about Feature-engine:

• Feature-engine: A new open-source Python package for feature engineering

• Practical Code Implementations of Feature Engineering for Machine Learning with Python

Documentation

• Documentation

En Español:

• Ingeniería de variables, MachinLenin, charla online