Linear functional regression with truncated signatures - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Journal of Multivariate Analysis Année : 2022

Linear functional regression with truncated signatures

Résumé

We place ourselves in a functional regression setting and propose a novel methodology for regressing a real output on vector-valued functional covariates. This methodology is based on the notion of signature, which is a representation of a function as an infinite series of its iterated integrals. The signature depends crucially on a truncation parameter for which an estimator is provided, together with theoretical guarantees. The complete procedure is tested on real-world datasets.

Dates et versions

hal-02869702 , version 1 (16-06-2020)

Identifiants

Citer

Adeline Fermanian. Linear functional regression with truncated signatures. Journal of Multivariate Analysis, 2022, 192, pp.105031. ⟨10.1016/j.jmva.2022.105031⟩. ⟨hal-02869702⟩
34 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More