Getting Started

scikit-feature is an open-source feature selection repository in Python, originally developed by the Data Mining and Machine Learning (DMML) Lab at Arizona State University. This fork maintains and updates the library for compatibility with modern versions of scikit-learn.

What is feature selection?

Feature selection is the process of selecting a subset of relevant features (variables, predictors) for use in model construction. It helps to:

• Reduce overfitting
• Improve accuracy
• Reduce training time
• Improve model interpretability

Algorithm categories

The library organizes feature selection algorithms into several categories:

Category	Description
Information Theoretical	Algorithms based on information theory (mutual information, entropy)
Similarity Based	Algorithms that use similarity measures between samples
Sparse Learning Based	Algorithms using sparse representation and l2,1-norm regularization
Statistical Based	Statistical methods for evaluating feature relevance
Structure Based	Algorithms leveraging data structure (graphs, groups, trees)
Wrapper Based	Algorithm selection using a predetermined classifier
Streaming	Online/streaming feature selection algorithms

Project structure

The single package lives in skfeature/, organized by algorithm family:

skfeature/
├── function/
│   ├── information_theoretical_based/  # Information-theoretic methods
│   ├── similarity_based/               # Similarity-based methods
│   ├── sparse_learning_based/          # Sparse learning methods
│   ├── statistical_based/              # Statistical methods
│   ├── structure/                      # Structure-based methods
│   ├── wrapper/                        # Wrapper methods
│   └── streaming/                      # Streaming methods
└── utility/                            # Utility functions

The Algorithms overview page lists every algorithm with a link to its reference documentation.

Next steps

• Installation — install the package with pip or from source
• Quick Start — run your first feature selection in minutes
• Algorithms overview — browse all available methods
• Contributing — learn how to add new algorithms

License

This project is licensed under the GNU General Public License v2.0.