Mapie

Mapie is a Python package compatible with Scikit-learn. It provides conformal prediction methods for quantifying uncertainty in regression and classification models. It uses a model-agnostic approach, making it easy to wrap around Scikit-learn models. The package supports prediction intervals and sets with theoretical guarantees through techniques like split conformal prediction and cross-validation. Mapie is applied in fields such as healthcare, finance, and engineering, where reliable predictions and risk management are critical. It is particularly useful for presenting confidence intervals in diagnostics and risk assessments in financial markets.

基本情報

公式サイト https://mapie.readthedocs.io/
公開度 ★★★
マニュアルページ https://mapie.readthedocs.io/
ダウンロードページ https://github.com/scikit-learn-contrib/MAPIE
ライセンス

3-clause BSD License

開発者
  • Main developers:
    • Development lead
      • Thibault Cordier
      • Vincent Blot
      • Louis Lacombe
    • Emeritus Core Developers
      • Grégoire Martinon
      • Vianney Taquet

For more details, please refer to the following GitHub page:

https://github.com/scikit-learn-contrib/MAPIE/blob/master/AUTHORS.rst#development-lead

対応OS・利用環境

python

関連サイト

[GitHub repository]

[Release note]