English

Estimating the conditional distribution in functional regression problems

Statistics Theory 2021-05-05 v1 Statistics Theory

Abstract

We consider the problem of consistently estimating the conditional distribution P(YAX)P(Y \in A |X) of a functional data object Y=(Y(t):t[0,1])Y=(Y(t): t\in[0,1]) given covariates XX in a general space, assuming that YY and XX are related by a functional linear regression model. Two natural estimation methods are proposed, based on either bootstrapping the estimated model residuals, or fitting functional parametric models to the model residuals and estimating P(YAX)P(Y \in A |X) via simulation. Whether either of these methods lead to consistent estimation depends on the consistency properties of the regression operator estimator, and the space within which YY is viewed. We show that under general consistency conditions on the regression operator estimator, which hold for certain functional principal component based estimators, consistent estimation of the conditional distribution can be achieved, both when YY is an element of a separable Hilbert space, and when YY is an element of the Banach space of continuous functions. The latter results imply that sets AA that specify path properties of YY, which are of interest in applications, can be considered. The proposed methods are studied in several simulation experiments, and data analyses of electricity price and pollution curves.

Keywords

Cite

@article{arxiv.2105.01412,
  title  = {Estimating the conditional distribution in functional regression problems},
  author = {Siegfried Hörmann and Thomas Kuenzer and Gregory Rice},
  journal= {arXiv preprint arXiv:2105.01412},
  year   = {2021}
}