Conformal inference for random objects

TL;DR

This paper introduces a new conformal inference method for object-valued responses in metric spaces, enabling accurate prediction sets with theoretical guarantees and practical applications to network and compositional data.

Contribution

It develops a novel conformal inference framework using conditional profile average transport costs for metric space-valued data, with proven asymptotic validity and superior finite-sample performance.

Findings

Outperforms existing methods in coverage and size of prediction sets

Provides asymptotic conditional validity of the proposed prediction sets

Demonstrates practical utility on network and compositional data

Abstract

We develop an inferential toolkit for analyzing object-valued responses, which correspond to data situated in general metric spaces, paired with Euclidean predictors within the conformal framework. To this end we introduce conditional profile average transport costs, where we compare distance profiles that correspond to one-dimensional distributions of probability mass falling into balls of increasing radius through the optimal transport cost when moving from one distance profile to another. The average transport cost to transport a given distance profile to all others is crucial for statistical inference in metric spaces and underpins the proposed conditional profile scores. A key feature of the proposed approach is to utilize the distribution of conditional profile average transport costs as conformity score for general metric space-valued responses, which facilitates the construction of prediction sets by the split conformal algorithm. We derive the uniform convergence rate of the proposed conformity score estimators and establish asymptotic conditional validity for the prediction sets. The finite sample performance for synthetic data in various metric spaces demonstrates that the proposed conditional profile score outperforms existing methods in terms of both coverage level and size of the resulting prediction sets, even in the special case of scalar Euclidean responses. We also demonstrate the practical utility of conditional profile scores for network data from New York taxi trips and for compositional data reflecting energy sourcing of U.S. states.