ET-COME

Equilibrium Transport with Conformal Minority Expansion

Welcome to ET-COME’s documentation! This project presents a novel ensemble method for handling class imbalance through equilibrium transport with conformal minority expansion.

Quick Links:

• Getting Started — Installation and quick start guide

• API Documentation — Complete API reference

• Release Notes — Version history and changelog

• References — Research papers and citations

• Additional Notes — Implementation details and notes

Key Features:

• Epistemic uncertainty identification for admissibility

• Risk-targeted optimal transport for synthesis placement

• OOB-based conformal screening for consistency

• Ensemble learning approach for imbalanced classification

• Comprehensive experimental framework

• Publication-ready implementation

About the Paper:

ET-COME is an ensemble method designed to address class imbalance in machine learning. It combines optimal transport with conformal prediction to generate synthetic minority samples in regions where the ensemble is genuinely uncertain.

For detailed information about the method, results, and experiments, please refer to the accompanying research paper and the Additional Notes section.