P. Raña Míguez, G. Aneiros, J. Vilar Fernández, P. Vieu
This study proposes naïve and wild bootstrap procedures to construct pointwise confidence intervals for two functional regression models. Specifically, we deal with the functional nonparametric regression model, considering scalar response and functional predictor, and also with the semi-functional partial linear regression model, in which we add linear effect of scalar covariates. By means of these two bootstrap procedures we can approximate the asymptotic distribution of the estimators in both regression models. The validity of these two methods has been proved theoretically in the setting of dependent data, assuming alpha-mixing conditions on the sample, and they were used to construct pointwise confidence intervals for each component of the functional regression models. A simulation study was carried out to show the performance of the proposed procedures. Applications to electrical data from the Spanish Electricity Market illustrate its usefulness in practice.
Palabras clave: Bootstrap, Confidence Intervals, Functional Data, Nonparametric regression, Partial linear regression, Dependent Data.Programado
L08.6 Sesión especial de Ramiro Melendreras
5 de septiembre de 2016 15:40
Paraninfo Envases de Cartón