Reducing statistical time-series problems to kernel-based binary classification
Résumé
It is shown how kernel-based methods developed for solving problems concerning i.i.d. data can be used for statistical analysis of highly-dependent time series. Specifically, the problems of time-series clustering, homogeneity testing and classification are addressed. The methods are based on a new distance between time-series distributions.
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|