Inferential procedures for partially observed functional data

J 2019

Inferential procedures for partially observed functional data

KRAUS, David

Základní údaje

Originální název

Inferential procedures for partially observed functional data

Autoři

KRAUS, David (203 Česká republika, garant, domácí)

Vydání

Journal of Multivariate Analysis, San Diego, Elsevier, 2019, 0047-259X

Další údaje

Jazyk

angličtina

Typ výsledku

Článek v odborném periodiku

Obor

10103 Statistics and probability

Stát vydavatele

Spojené státy

Utajení

není předmětem státního či obchodního tajemství

Odkazy

Full Text

Impakt faktor

Impact factor: 1.136

Kód RIV

RIV/00216224:14310/19:00107420

Organizační jednotka

Přírodovědecká fakulta

DOI

http://dx.doi.org/10.1016/j.jmva.2019.05.002

UT WoS

000481565500034

EID Scopus

2-s2.0-85066395567

Klíčová slova anglicky

Bootstrap; Covariance operator; Functional data; K-sample test; Partial observation; Principal components

Štítky

rivok

Příznaky

Mezinárodní význam, Recenzováno

Změněno: 18. 3. 2020 14:07, Mgr. Marie Novosadová Šípková, DiS.

Anotace

V originále

In functional data analysis it is usually assumed that all functions are completely, densely or sparsely observed on the same domain. Recent applications have brought attention to situations where each functional variable may be observed only on a subset of the domain while no information about the function is available on the complement. Various advanced methods for such partially observed functional data have already been developed but, interestingly, some essential methods, such as K-sample tests of equal means or covariances and confidence intervals for eigenvalues and eigenfunctions, are lacking. Without requiring any complete curves in the data, we derive asymptotic distributions of estimators of the mean function, covariance operator and eigenelements and construct hypothesis tests and confidence intervals. To overcome practical difficulties with storing large objects in computer memory, which arise due to partial observation, we use the nonparametric bootstrap approach. The proposed methods are investigated theoretically, in simulations and on a fragmentary functional data set from medical research.

Návaznosti

GJ17-22950Y, projekt VaV

Název: Statistická inference pro složité náhodné procesy v ekonometrickém modelování

Investor: Grantová agentura ČR, Statistická inference pro složité náhodné procesy v ekonometrickém modelování

Citovat

KRAUS, David. Inferential procedures for partially observed functional data. Journal of Multivariate Analysis. San Diego: Elsevier, 2019, roč. 173, September, s. 583-603. ISSN 0047-259X. Dostupné z: https://dx.doi.org/10.1016/j.jmva.2019.05.002.

@article{1534363,
   author = {Kraus, David},
   article_location = {San Diego},
   article_number = {September},
   doi = {http://dx.doi.org/10.1016/j.jmva.2019.05.002},
   keywords = {Bootstrap; Covariance operator; Functional data; K-sample test; Partial observation; Principal components},
   language = {eng},
   issn = {0047-259X},
   journal = {Journal of Multivariate Analysis},
   title = {Inferential procedures for partially observed functional data},
   url = {https://www.sciencedirect.com/science/article/pii/S0047259X18304950},
   volume = {173},
   year = {2019}
}

TY  - JOUR
ID  - 1534363
AU  - Kraus, David
PY  - 2019
TI  - Inferential procedures for partially observed functional data
JF  - Journal of Multivariate Analysis
VL  - 173
IS  - September
SP  - 583-603
EP  - 583-603
PB  - Elsevier
SN  - 0047259X
KW  - Bootstrap
KW  - Covariance operator
KW  - Functional data
KW  - K-sample test
KW  - Partial observation
KW  - Principal components
UR  - https://www.sciencedirect.com/science/article/pii/S0047259X18304950
L2  - https://www.sciencedirect.com/science/article/pii/S0047259X18304950
N2  - In functional data analysis it is usually assumed that all functions are completely, densely or sparsely observed on the same domain. Recent applications have brought attention to situations where each functional variable may be observed only on a subset of the domain while no information about the function is available on the complement. Various advanced methods for such partially observed functional data have already been developed but, interestingly, some essential methods, such as K-sample tests of equal means or covariances and confidence intervals for eigenvalues and eigenfunctions, are lacking. Without requiring any complete curves in the data, we derive asymptotic distributions of estimators of the mean function, covariance operator and eigenelements and construct hypothesis tests and confidence intervals. To overcome practical difficulties with storing large objects in computer memory, which arise due to partial observation, we use the nonparametric bootstrap approach. The proposed methods are investigated theoretically, in simulations and on a fragmentary functional data set from medical research.
ER  -

KRAUS, David. Inferential procedures for partially observed functional data. \textit{Journal of Multivariate Analysis}. San Diego: Elsevier, 2019, roč.~173, September, s.~583-603. ISSN~0047-259X. Dostupné z: https://dx.doi.org/10.1016/j.jmva.2019.05.002.

Přehled o publikaci