Stochastic Online Convex Optimization; Application to probabilistic time series forecasting - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2022

Stochastic Online Convex Optimization; Application to probabilistic time series forecasting

Résumé

We present a general approach for obtaining stochastic regret bounds for non-convex loss functions holding with high probability. Scale-free algorithms can solve Stochastic Online Convex Optimization using "surrogate losses" regret analysis. Then, we provide optimal prediction and probabilistic forecasting methods for non-stationary unbounded time series.
Fichier principal
Vignette du fichier
SOCO.pdf (592.67 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03125863 , version 1 (29-01-2021)
hal-03125863 , version 2 (26-04-2021)
hal-03125863 , version 3 (25-02-2022)
hal-03125863 , version 4 (31-03-2023)

Identifiants

Citer

Olivier Wintenberger. Stochastic Online Convex Optimization; Application to probabilistic time series forecasting. 2022. ⟨hal-03125863v3⟩
150 Consultations
170 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More