Functional linear model with missing values in the covariate and the response - Institut de Mathématiques et de Modélisation de Montpellier Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2020

Functional linear model with missing values in the covariate and the response

Résumé

Dealing with missing values is an important issue in data observation or data recording process. In this paper, we consider a functional linear regression model, where some observations of the real response and the functional covariate are affected by missing data. We use a reconstruction operator that aims to recover the missing parts of the explanatory curves, then we are interested in regression imputation method of missing data on the response variable, using functional principal component regression to estimate the functional coefficient of the model. We study the asymptotic behavior of the prediction error we commit when missing data are replaced by the imputed values in the original dataset. The practical behavior of the method is also studied on simulated data and a real dataset.
Fichier principal
Vignette du fichier
Article_Crambesetal2020_jcg_template.pdf (688.21 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03083293 , version 1 (18-12-2020)
hal-03083293 , version 2 (10-09-2021)
hal-03083293 , version 3 (30-05-2022)

Identifiants

  • HAL Id : hal-03083293 , version 1

Citer

Christophe Crambes, Chayma Daayeb, Ali Gannoun, Yousri Henchiri. Functional linear model with missing values in the covariate and the response. 2020. ⟨hal-03083293v1⟩
228 Consultations
188 Téléchargements

Partager

Gmail Facebook X LinkedIn More