A Binary-Classification-Based Metric between Time-Series Distributions and Its Use in Statistical and Learning Problems

Daniil Ryabko 1 Jérémie Mary 1
1 SEQUEL - Sequential Learning
LIFL - Laboratoire d'Informatique Fondamentale de Lille, LAGIS - Laboratoire d'Automatique, Génie Informatique et Signal, Inria Lille - Nord Europe
Abstract : A metric between time-series distributions is proposed that can be evaluated using binary classification methods, which were originally developed to work on i.i.d.\ data. It is shown how this metric can be used for solving statistical problems that are seemingly unrelated to classification and concern highly dependent time series. Specifically, the problems of time-series clustering, homogeneity testing and the three-sample problem are addressed. Universal consistency of the resulting algorithms is proven under most general assumptions. The theoretical results are illustrated with experiments on synthetic and real-world data.
Type de document :
Article dans une revue
Journal of Machine Learning Research, Journal of Machine Learning Research, 2013, 14, pp.2837-2856
Liste complète des métadonnées

https://hal.inria.fr/hal-00913240
Contributeur : Daniil Ryabko <>
Soumis le : mardi 3 décembre 2013 - 14:24:57
Dernière modification le : mercredi 31 janvier 2018 - 17:31:32

Identifiants

  • HAL Id : hal-00913240, version 1

Citation

Daniil Ryabko, Jérémie Mary. A Binary-Classification-Based Metric between Time-Series Distributions and Its Use in Statistical and Learning Problems. Journal of Machine Learning Research, Journal of Machine Learning Research, 2013, 14, pp.2837-2856. 〈hal-00913240〉

Partager

Métriques

Consultations de la notice

288