Consistency of the mean and the principal components of spatially distributed functional data

Siegfried Hörmann, P. Kokoszka

Research output: Contribution to journalArticle

Abstract

This paper develops a framework for the estimation of the functional mean and the functional principal components when the functions form a random field. We establish conditions for the sample average (in space) to be a consistent estimator of the population mean function, and for the usual empirical covariance operator to be a consistent estimator of the population covariance operator. These conditions involve an interplay of the assumptions on an appropriately defined dependence between the random functions and the assumptions on the spatial distribution of the sampling points. The rates of convergence may be the same as for iid functional samples, but generally depend on the strength of dependence and appropriately quantified distances between the sampling points. We also formulate conditions for the lack of consistency
LanguageUndefined/Unknown
Pages1535-1558
Number of pages24
JournalBernoulli
Volume19
StatusPublished - 2013

Cite this

Consistency of the mean and the principal components of spatially distributed functional data. / Hörmann, Siegfried; Kokoszka, P.

In: Bernoulli, Vol. 19, 2013, p. 1535-1558.

Research output: Contribution to journalArticle

@article{e68d0d63a2204a77b10d7f9072981f4c,
title = "Consistency of the mean and the principal components of spatially distributed functional data",
abstract = "This paper develops a framework for the estimation of the functional mean and the functional principal components when the functions form a random field. We establish conditions for the sample average (in space) to be a consistent estimator of the population mean function, and for the usual empirical covariance operator to be a consistent estimator of the population covariance operator. These conditions involve an interplay of the assumptions on an appropriately defined dependence between the random functions and the assumptions on the spatial distribution of the sampling points. The rates of convergence may be the same as for iid functional samples, but generally depend on the strength of dependence and appropriately quantified distances between the sampling points. We also formulate conditions for the lack of consistency",
author = "Siegfried H{\"o}rmann and P. Kokoszka",
note = "Language of publication: en",
year = "2013",
language = "undefiniert/unbekannt",
volume = "19",
pages = "1535--1558",
journal = "Bernoulli",
issn = "1350-7265",
publisher = "International Statistical Institute",

}

TY - JOUR

T1 - Consistency of the mean and the principal components of spatially distributed functional data

AU - Hörmann,Siegfried

AU - Kokoszka,P.

N1 - Language of publication: en

PY - 2013

Y1 - 2013

N2 - This paper develops a framework for the estimation of the functional mean and the functional principal components when the functions form a random field. We establish conditions for the sample average (in space) to be a consistent estimator of the population mean function, and for the usual empirical covariance operator to be a consistent estimator of the population covariance operator. These conditions involve an interplay of the assumptions on an appropriately defined dependence between the random functions and the assumptions on the spatial distribution of the sampling points. The rates of convergence may be the same as for iid functional samples, but generally depend on the strength of dependence and appropriately quantified distances between the sampling points. We also formulate conditions for the lack of consistency

AB - This paper develops a framework for the estimation of the functional mean and the functional principal components when the functions form a random field. We establish conditions for the sample average (in space) to be a consistent estimator of the population mean function, and for the usual empirical covariance operator to be a consistent estimator of the population covariance operator. These conditions involve an interplay of the assumptions on an appropriately defined dependence between the random functions and the assumptions on the spatial distribution of the sampling points. The rates of convergence may be the same as for iid functional samples, but generally depend on the strength of dependence and appropriately quantified distances between the sampling points. We also formulate conditions for the lack of consistency

M3 - Artikel

VL - 19

SP - 1535

EP - 1558

JO - Bernoulli

T2 - Bernoulli

JF - Bernoulli

SN - 1350-7265

ER -