Minimax-rate adaptive nonparametric regression with unknown correlations of errors

Guowu Yang, Yuhong Yang

Research output: Contribution to journalArticle

Abstract

Minimax-rate adaptive nonparametric regression has been intensively studied under the assumption of independent or uncorrelated errors in the literature. In many applications, however, the errors are dependent, including both short- and long-range dependent situations. In such a case, adaptation with respect to the unknown dependence is important. We present a general result in this direction under Gaussian errors. It is assumed that the covariance matrix of the errors is known to be in a list of specifications possibly including independence, short-range dependence and long-range dependence as well. The regression function is known to be in a countable (or uncountable but well-structured) collection of function classes. Adaptive estimators are constructed to attain the minimax rate of convergence automatically for each function class under each correlation specification in the corresponding lists.

Original languageEnglish (US)
Pages (from-to)227-244
Number of pages18
JournalScience China Mathematics
Volume62
Issue number2
DOIs
StatePublished - Feb 1 2019

Fingerprint

Minimax Rate
Nonparametric Regression
Unknown
Range of data
Specification
Adaptive Estimator
Long-range Dependence
Dependent
Uncountable
Regression Function
Covariance matrix
Countable
Rate of Convergence
Class

Keywords

  • 62C20
  • 62G08
  • adaptive estimation
  • long-range dependence
  • nonparametric regression
  • rate of convergence

Cite this

Minimax-rate adaptive nonparametric regression with unknown correlations of errors. / Yang, Guowu; Yang, Yuhong.

In: Science China Mathematics, Vol. 62, No. 2, 01.02.2019, p. 227-244.

Research output: Contribution to journalArticle

@article{ffe7dab09e524b72a5cb98cdd38b02d8,
title = "Minimax-rate adaptive nonparametric regression with unknown correlations of errors",
abstract = "Minimax-rate adaptive nonparametric regression has been intensively studied under the assumption of independent or uncorrelated errors in the literature. In many applications, however, the errors are dependent, including both short- and long-range dependent situations. In such a case, adaptation with respect to the unknown dependence is important. We present a general result in this direction under Gaussian errors. It is assumed that the covariance matrix of the errors is known to be in a list of specifications possibly including independence, short-range dependence and long-range dependence as well. The regression function is known to be in a countable (or uncountable but well-structured) collection of function classes. Adaptive estimators are constructed to attain the minimax rate of convergence automatically for each function class under each correlation specification in the corresponding lists.",
keywords = "62C20, 62G08, adaptive estimation, long-range dependence, nonparametric regression, rate of convergence",
author = "Guowu Yang and Yuhong Yang",
year = "2019",
month = "2",
day = "1",
doi = "10.1007/s11425-018-9394-x",
language = "English (US)",
volume = "62",
pages = "227--244",
journal = "Science China Mathematics",
issn = "1674-7283",
publisher = "Science in China Press",
number = "2",

}

TY - JOUR

T1 - Minimax-rate adaptive nonparametric regression with unknown correlations of errors

AU - Yang, Guowu

AU - Yang, Yuhong

PY - 2019/2/1

Y1 - 2019/2/1

N2 - Minimax-rate adaptive nonparametric regression has been intensively studied under the assumption of independent or uncorrelated errors in the literature. In many applications, however, the errors are dependent, including both short- and long-range dependent situations. In such a case, adaptation with respect to the unknown dependence is important. We present a general result in this direction under Gaussian errors. It is assumed that the covariance matrix of the errors is known to be in a list of specifications possibly including independence, short-range dependence and long-range dependence as well. The regression function is known to be in a countable (or uncountable but well-structured) collection of function classes. Adaptive estimators are constructed to attain the minimax rate of convergence automatically for each function class under each correlation specification in the corresponding lists.

AB - Minimax-rate adaptive nonparametric regression has been intensively studied under the assumption of independent or uncorrelated errors in the literature. In many applications, however, the errors are dependent, including both short- and long-range dependent situations. In such a case, adaptation with respect to the unknown dependence is important. We present a general result in this direction under Gaussian errors. It is assumed that the covariance matrix of the errors is known to be in a list of specifications possibly including independence, short-range dependence and long-range dependence as well. The regression function is known to be in a countable (or uncountable but well-structured) collection of function classes. Adaptive estimators are constructed to attain the minimax rate of convergence automatically for each function class under each correlation specification in the corresponding lists.

KW - 62C20

KW - 62G08

KW - adaptive estimation

KW - long-range dependence

KW - nonparametric regression

KW - rate of convergence

UR - http://www.scopus.com/inward/record.url?scp=85060235621&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85060235621&partnerID=8YFLogxK

U2 - 10.1007/s11425-018-9394-x

DO - 10.1007/s11425-018-9394-x

M3 - Article

AN - SCOPUS:85060235621

VL - 62

SP - 227

EP - 244

JO - Science China Mathematics

JF - Science China Mathematics

SN - 1674-7283

IS - 2

ER -