Conditional validity of inductive conformal predictors
Machine Learning
2012-09-25 v2
Abstract
Conformal predictors are set predictors that are automatically valid in the sense of having coverage probability equal to or exceeding a given confidence level. Inductive conformal predictors are a computationally efficient version of conformal predictors satisfying the same property of validity. However, inductive conformal predictors have been only known to control unconditional coverage probability. This paper explores various versions of conditional validity and various ways to achieve them using inductive conformal predictors and their modifications.
Keywords
Cite
@article{arxiv.1209.2673,
title = {Conditional validity of inductive conformal predictors},
author = {Vladimir Vovk},
journal= {arXiv preprint arXiv:1209.2673},
year = {2012}
}
Comments
23 pages, 9 figures, 2 tables; to appear in the ACML 2012 Proceedings