A Binary-Classification-Based Metric between Time-Series Distributions and Its Use in Statistical and Learning Problems - Archive ouverte HAL Access content directly
Journal Articles Journal of Machine Learning Research Year : 2013

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

(1) , (1)
1
Jérémie Mary

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.
Not file

Dates and versions

hal-00913240 , version 1 (03-12-2013)

Identifiers

  • HAL Id : hal-00913240 , version 1

Cite

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, 2013, 14, pp.2837-2856. ⟨hal-00913240⟩
216 View
0 Download

Share

Gmail Facebook Twitter LinkedIn More