Conformal Prediction Regions are Imprecise Highest Density Regions

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Caprio, Michele, Sale, Yusuf, Hüllermeier, Eyke
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916695444029440
author Caprio, Michele
Sale, Yusuf
Hüllermeier, Eyke
author_facet Caprio, Michele
Sale, Yusuf
Hüllermeier, Eyke
contents Recently, Cella and Martin proved how, under an assumption called consonance, a credal set (i.e. a closed and convex set of probabilities) can be derived from the conformal transducer associated with transductive conformal prediction. We show that the Imprecise Highest Density Region (IHDR) associated with such a credal set corresponds to the classical Conformal Prediction Region. In proving this result, we establish a new relationship between Conformal Prediction and Imprecise Probability (IP) theories, via the IP concept of a cloud. A byproduct of our presentation is the discovery that consonant plausibility functions are monoid homomorphisms, a new algebraic property of an IP tool.
format Preprint
id arxiv_https___arxiv_org_abs_2502_06331
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Conformal Prediction Regions are Imprecise Highest Density Regions
Caprio, Michele
Sale, Yusuf
Hüllermeier, Eyke
Machine Learning
Probability
Primary: 68T37, Secondary: 62M20, 60G25, 20M32, 15A80
Recently, Cella and Martin proved how, under an assumption called consonance, a credal set (i.e. a closed and convex set of probabilities) can be derived from the conformal transducer associated with transductive conformal prediction. We show that the Imprecise Highest Density Region (IHDR) associated with such a credal set corresponds to the classical Conformal Prediction Region. In proving this result, we establish a new relationship between Conformal Prediction and Imprecise Probability (IP) theories, via the IP concept of a cloud. A byproduct of our presentation is the discovery that consonant plausibility functions are monoid homomorphisms, a new algebraic property of an IP tool.
title Conformal Prediction Regions are Imprecise Highest Density Regions
topic Machine Learning
Probability
Primary: 68T37, Secondary: 62M20, 60G25, 20M32, 15A80
url https://arxiv.org/abs/2502.06331