On studentising and blocklength selection for the bootstrap on time series

M Peifer, B Schelter, B Guschlbauer, B Hellwig, C H Lucking, J Timmer

Research output: Contribution to journalArticle

8 Citations (Scopus)

Abstract

For independent data, non-parametric bootstrap is realised by resampling the data with replacement. This approach fails for dependent data such as time series. If the data generating process is at least stationary and mixing, the blockwise bootstrap by drawing subsamples or blocks of the data saves the concept. For the blockwise bootstrap a blocklength has to be selected. We propose a method for selecting the optimal blocklength. To improve the finite size properties of the blockwise bootstrap, studentised statistics is considered. If the statistic can be represented as a smooth function model this studentisation can be approximated efficiently. The studentised blockwise bootstrap method is applied for testing hypotheses on medical time series.

Original languageEnglish
Pages (from-to)346-357
Number of pages12
JournalBiometrical Journal
Volume47
Issue number3
DOIs
Publication statusPublished - Jun 2005

Keywords

  • blockwise bootstrap
  • studentising
  • smooth function model
  • tremor

Cite this

Peifer, M., Schelter, B., Guschlbauer, B., Hellwig, B., Lucking, C. H., & Timmer, J. (2005). On studentising and blocklength selection for the bootstrap on time series. Biometrical Journal, 47(3), 346-357. https://doi.org/10.1002/bimj.200310112

On studentising and blocklength selection for the bootstrap on time series. / Peifer, M ; Schelter, B ; Guschlbauer, B ; Hellwig, B ; Lucking, C H ; Timmer, J .

In: Biometrical Journal, Vol. 47, No. 3, 06.2005, p. 346-357.

Research output: Contribution to journalArticle

Peifer, M, Schelter, B, Guschlbauer, B, Hellwig, B, Lucking, CH & Timmer, J 2005, 'On studentising and blocklength selection for the bootstrap on time series', Biometrical Journal, vol. 47, no. 3, pp. 346-357. https://doi.org/10.1002/bimj.200310112
Peifer, M ; Schelter, B ; Guschlbauer, B ; Hellwig, B ; Lucking, C H ; Timmer, J . / On studentising and blocklength selection for the bootstrap on time series. In: Biometrical Journal. 2005 ; Vol. 47, No. 3. pp. 346-357.
@article{735c9b1a2f2040d0b9e1d8c8fac29c6c,
title = "On studentising and blocklength selection for the bootstrap on time series",
abstract = "For independent data, non-parametric bootstrap is realised by resampling the data with replacement. This approach fails for dependent data such as time series. If the data generating process is at least stationary and mixing, the blockwise bootstrap by drawing subsamples or blocks of the data saves the concept. For the blockwise bootstrap a blocklength has to be selected. We propose a method for selecting the optimal blocklength. To improve the finite size properties of the blockwise bootstrap, studentised statistics is considered. If the statistic can be represented as a smooth function model this studentisation can be approximated efficiently. The studentised blockwise bootstrap method is applied for testing hypotheses on medical time series.",
keywords = "blockwise bootstrap, studentising, smooth function model, tremor",
author = "M Peifer and B Schelter and B Guschlbauer and B Hellwig and Lucking, {C H} and J Timmer",
year = "2005",
month = "6",
doi = "10.1002/bimj.200310112",
language = "English",
volume = "47",
pages = "346--357",
journal = "Biometrical Journal",
issn = "0323-3847",
publisher = "Wiley-VCH Verlag",
number = "3",

}

TY - JOUR

T1 - On studentising and blocklength selection for the bootstrap on time series

AU - Peifer, M

AU - Schelter, B

AU - Guschlbauer, B

AU - Hellwig, B

AU - Lucking, C H

AU - Timmer, J

PY - 2005/6

Y1 - 2005/6

N2 - For independent data, non-parametric bootstrap is realised by resampling the data with replacement. This approach fails for dependent data such as time series. If the data generating process is at least stationary and mixing, the blockwise bootstrap by drawing subsamples or blocks of the data saves the concept. For the blockwise bootstrap a blocklength has to be selected. We propose a method for selecting the optimal blocklength. To improve the finite size properties of the blockwise bootstrap, studentised statistics is considered. If the statistic can be represented as a smooth function model this studentisation can be approximated efficiently. The studentised blockwise bootstrap method is applied for testing hypotheses on medical time series.

AB - For independent data, non-parametric bootstrap is realised by resampling the data with replacement. This approach fails for dependent data such as time series. If the data generating process is at least stationary and mixing, the blockwise bootstrap by drawing subsamples or blocks of the data saves the concept. For the blockwise bootstrap a blocklength has to be selected. We propose a method for selecting the optimal blocklength. To improve the finite size properties of the blockwise bootstrap, studentised statistics is considered. If the statistic can be represented as a smooth function model this studentisation can be approximated efficiently. The studentised blockwise bootstrap method is applied for testing hypotheses on medical time series.

KW - blockwise bootstrap

KW - studentising

KW - smooth function model

KW - tremor

U2 - 10.1002/bimj.200310112

DO - 10.1002/bimj.200310112

M3 - Article

VL - 47

SP - 346

EP - 357

JO - Biometrical Journal

JF - Biometrical Journal

SN - 0323-3847

IS - 3

ER -