EUVIMEDEUVIMED entdecken
Crossref

Training-conditional coverage for distribution-free predictive inference

Michael Bian, Rina Foygel Barber

Electronic Journal of Statistics · 01.01.2023 · 17 (2) · 2044-2066

DOI: 10.1214/23-ejs2145 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Electronic Journal of Statistics, erschienen am 01.01.2023. Verfasst von Michael Bian und 1 weiteren Autor:innen. Die Arbeit wurde bisher 9-mal zitiert (Crossref) und nennt 25 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The field of distribution-free predictive inference provides tools for provably valid prediction without any assumptions on the distribution of the data, which can be paired with any regression algorithm to provide accurate and reliable predictive intervals. The guarantees provided by these methods are typically marginal, meaning that predictive accuracy holds on average over both the training data set and the test point that is queried. However, it may be preferable to obtain a stronger guarantee of training-conditional coverage, which would ensure that most draws of the training data set result in accurate predictive accuracy on future test points. This property is known to hold for the split conformal prediction method. In this work, we examine the training-conditional coverage properties of several other distribution-free predictive inference methods, and find that training-conditional coverage is achieved by some methods but is impossible to guarantee without further assumptions for others.

Zitieren

APA

Bian, M., & Barber, R. F. (2023). Training-conditional coverage for distribution-free predictive inference. Electronic Journal of Statistics, 17(2), 2044-2066. https://doi.org/10.1214/23-ejs2145

Vancouver

Bian M, Barber RF. Training-conditional coverage for distribution-free predictive inference. Electronic Journal of Statistics. 2023;17(2):2044-2066. doi:10.1214/23-ejs2145

BibTeX
@article{Bian2023,
  author = {Bian, Michael and Barber, Rina Foygel},
  title = {Training-conditional coverage for distribution-free predictive inference},
  journal = {Electronic Journal of Statistics},
  year = {2023},
  volume = {17},
  number = {2},
  pages = {2044--2066},
  doi = {10.1214/23-ejs2145}
}
Kostenlos starten →