Shapley
A Python library for evaluating binary classifiers in machine learning ensembles using Shapley value computation and approximation methods.
What is Shapley?
Shapley is a Python library for evaluating binary classifiers within machine learning ensembles by computing their Shapley values in weighted voting games. It quantifies each model's contribution to the ensemble's performance and assesses player pool heterogeneity through Shapley entropy. The library implements rigorous methods from economics and computer science to provide a robust framework for classifier evaluation.
Machine learning researchers and data scientists working with ensemble models who need to analyze individual classifier contributions and ensemble diversity. It is particularly suited for those implementing or studying cooperative game theory applications in ML.
Developers choose Shapley for its research-backed implementation of multiple Shapley value computation methods, including exact and approximation techniques, in a single, well-documented library. Its unique focus on weighted voting games and Shapley entropy for heterogeneity analysis sets it apart from general-purpose Shapley value tools.
Overview
The official implementation of "The Shapley Value of Classifiers in Ensemble Games" (CIKM 2021).
Use Cases
Best For
- Quantifying the contribution of individual binary classifiers to an ensemble's predictive performance.
Related Projects
Found a gem we're missing?
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
