Conformal Prediction under Hypergraphical Models

Abstract : Conformal predictors are usually defined and studied under the exchangeability assumption. However, their definition can be extended to a wide class of statistical models, called online compression models, while retaining their property of automatic validity. This paper is devoted to conformal prediction under hypergraphical models that are more specific than the exchangeability model. Namely, we define two natural classes of conformity measures for such hypergraphical models and study the corresponding conformal predictors empirically on benchmark LED data sets. Our experiments show that they are more efficient than conformal predictors that use only the exchangeability assumption.
Type de document :
Communication dans un congrès
Harris Papadopoulos; Andreas S. Andreou; Lazaros Iliadis; Ilias Maglogiannis. 9th Artificial Intelligence Applications and Innovations (AIAI), Sep 2013, Paphos, Greece. Springer, IFIP Advances in Information and Communication Technology, AICT-412, pp.371-383, 2013, Artificial Intelligence Applications and Innovations. 〈10.1007/978-3-642-41142-7_38〉
Liste complète des métadonnées

Littérature citée [6 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-01459632
Contributeur : Hal Ifip <>
Soumis le : mardi 7 février 2017 - 13:05:42
Dernière modification le : vendredi 1 décembre 2017 - 01:16:34
Document(s) archivé(s) le : lundi 8 mai 2017 - 14:15:25

Fichier

978-3-642-41142-7_38_Chapter.p...
Fichiers produits par l'(les) auteur(s)

Licence


Distributed under a Creative Commons Paternité 4.0 International License

Identifiants

Citation

Valentina Fedorova, Alex Gammerman, Ilia Nouretdinov, Vladimir Vovk. Conformal Prediction under Hypergraphical Models. Harris Papadopoulos; Andreas S. Andreou; Lazaros Iliadis; Ilias Maglogiannis. 9th Artificial Intelligence Applications and Innovations (AIAI), Sep 2013, Paphos, Greece. Springer, IFIP Advances in Information and Communication Technology, AICT-412, pp.371-383, 2013, Artificial Intelligence Applications and Innovations. 〈10.1007/978-3-642-41142-7_38〉. 〈hal-01459632〉

Partager

Métriques

Consultations de la notice

377

Téléchargements de fichiers

21