Robust set estimation via conformal inference

A. Cholaquidis, E. Joly, L. Moreno

Conformal inference provides a general framework for constructing prediction
sets with finite-sample coverage guarantees without imposing parametric as-
sumptions on the data-generating distribution. However, classical conformal
methods are known to be highly sensitive to outliers and departures from the
idealized model assumptions.

In this talk, we present a robust conformal inference approach to set esti-
mation, based on nonconformity scores built from tools in robust statistics and
topological data analysis, such as multivariate depth functions and distance-to-
measure constructions.

Palabras clave: conformal inferences, set estimation

GT Estadística no Paramétrica I: Estimación no paramétrica
4 de septiembre de 2026  09:00
Aula 29

P. García Soidán, T. R. Cotos Yáñez

H. González-Vázquez, B. Pateiro-López, A. Rodríguez-Casal

J. E. Chacón, J. Fernández Serrano

N. Sánchez, M. Conde Amboage, M. I. Borrajo García