Copula-based conformal prediction for multi-target regression

作者：

Highlights：

• highlights

• We use a conformal prediction setting for Multi-Target Regression.

• We link copulas to the estimation of non-conformity scores.

• We perform experiments with 3 types of copulas, and compare between them.

• We show through experiments that our method performs better than a naive approach.

摘要

highlights•We use a conformal prediction setting for Multi-Target Regression.•We link copulas to the estimation of non-conformity scores.•We perform experiments with 3 types of copulas, and compare between them.•We show through experiments that our method performs better than a naive approach.

论文关键词：Inductive conformal prediction,Copula functions,Multi-target regression,Deep neural networks,Random forests

论文评审过程：Received 27 August 2020, Revised 2 April 2021, Accepted 5 May 2021, Available online 16 June 2021, Version of Record 1 July 2021.

论文官网地址：https://doi.org/10.1016/j.patcog.2021.108101