Hypergraph-based data mining tool for binary classification.
Project description
Hypper is a data-mining Python library for binary classification. It uses hypergraph-based methods to explore datasets for the purpose of undersampling, feature selection and binary classification.
Hypper provides an easy-to-use interface familiar to well-recognized Scikit-Learn API.
The primary goal of this library is to provide a tool for handling datasets consisting of mainly categorical features. Novel hypergraph-based methods proposed in the Hypper library were benchmarked against the alternative solutions and achieved satisfactory results. More details can be found in scientific papers presented in the section below.
Installation
Minimal installation
pip install hypper
Installation with documentation
pip install -e .['documentation']
Installation with tests
pip install -e .['testing']
Install everything
pip install -e .['all']
Usage
Undersampling
Feature Selection
Tutorials:
- Undersampling
- Feature Selection
- Classification
Testing
pytest
Important links
- Source code - https://github.com/sleter/hypper
- Documentation - https://hypper-team.github.io/hypper.html
Citation
Related articles
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
