Aggregation in conformal e-classification

2026-05-08

AI Summary

This study addresses the challenge of efficiently aggregating multiple conformal e-predictors while preserving the validity guarantees of conformal prediction and achieving a favorable trade-off between predictive performance and computational efficiency. To this end, the authors propose a streamlined and more flexible aggregation method based on a cross-aggregation strategy, which features conceptual clarity and straightforward implementation while outperforming conventional aggregation schemes. Experimental results demonstrate that the proposed approach significantly enhances prediction efficiency and computational scalability, all while maintaining approximate validity guarantees inherent to conformal prediction frameworks.

Abstract

Aggregating conformal predictors is a standard way of balancing their predictive and computational efficiency while retaining their validity, at least approximately. An important advantage of conformal e-predictors is that they are easier to aggregate without sacrificing their validity. This paper studies experimentally cross-conformal e-prediction, which is an existing method of aggregating conformal e-predictors, and its modifications that are conceptually simpler and more flexible.

Problem

• conformal prediction
• e-prediction
• aggregation
• validity
• computational efficiency

Innovation

• conformal e-prediction
• aggregation
• cross-conformal
• validity
• computational efficiency