zbMATH — the first resource for mathematics

Uniform consistency and uniform in bandwidth consistency for nonparametric regression estimates and conditional $$U$$-statistics involving functional data. (English) Zbl 07214281
Summary: W. Stute [(1991), Annals of Probability, 19, 812-825] introduced a class of so-called conditional $$U$$-statistics, which may be viewed as a generalisation of the Nadaraya-Watson estimates of a regression function. Stute proved their strong pointwise consistency to
$m(\mathbf{t} := \mathbb{E}[\varphi(Y_1, \dots, Y_m)|(X_q, \dots, X_m) = \mathbf{t}], \quad \text{for } \mathbf{t} \in \mathbb{R}^{dm}.$
We apply the methods developed in Dony and Mason [(2008), Bernoulli, 14(4), 1108-1133] to establish uniformity in $$\mathbf{t}$$ and in bandwidth consistency (i.e. $$h_n, h_n \ in [a_n, a_n]$$ where $$0 < a_n < b_n \to 0$$ at some specific rate) to $$m(\mathbf{t}$$ of the estimator proposed by Stute when $$Y$$ and covariates $$X$$ are functional taking value in some abstract spaces. In addition, uniform consistency is also established over $$\varphi \in \mathscr{F}$$ for a suitably restricted class $$\mathscr{F}$$. The theoretical uniform consistency results, established in this paper, are (or will be) key tools for many further developments in functional data analysis. Applications include the Nadaraya-Watson kernel estimators and the conditional distribution function. Our theorems allow data-driven local bandwidths for these statistics.
MSC:
 62G08 Nonparametric regression and quantile regression 62R10 Functional data analysis
fda (R)
Full Text: