Universal distribution of the empirical coverage in split conformal prediction

Paulo C. Marques F
3/5/2023
math.STcs.LGstat.ML

No official code found. Best implementation: paulocmarquesf/coverage (3 stars, last commit 2026-01-19).

Abstract

When split conformal prediction operates in batch mode with exchangeable data, we determine the exact distribution of the empirical coverage of prediction sets produced for a finite batch of future observables, as well as the exact distribution of its almost sure limit when the batch size goes to infinity. Both distributions are universal, being determined solely by the nominal miscoverage level and the calibration sample size, thereby establishing a criterion for choosing the minimum required calibration sample size in applications.

View on arXivView PDF

Code Implementations(1)

paulocmarquesf/coverageCommunity100%

Universal distribution of the empirical coverage in split conformal prediction

31RSep 21, 20249 months ago

Similar Papers

Split Conformal Prediction under Data Contamination

85% match2024

Aggregating Conformal Prediction Sets via α-Allocation

84% match2025

Probabilistic Conformal Coverage Guarantees in Small-Data Settings

84% match2025

On the Expected Size of Conformal Prediction Sets

84% match1 repo62023

Conformal Predictions under Markovian Data

83% match2024

Multi-Distribution Robust Conformal Prediction

82% match1 repo62026

Cite this paper

@article{f2023universal,
  title  = {Universal distribution of the empirical coverage in split conformal prediction},
  author = {Paulo C. Marques F},
  year   = {2023},
  eprint = {2303.02770},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2303.02770v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2303.02770.svg)](https://codeofpaper.com/paper/2303.02770)

Discussion