VARIANCE ESTIMATION FOR SAMPLE AUTOCOVARIANCES: DIRECT AND RESAMPLING APPROACHES

Clifford M. Hurvich, Jeffrey S. Simonoff, Scott Zeger

Research output: Contribution to journalArticle

Abstract

The usual covariance estimates for data n‐1 from a stationary zero‐mean stochastic process {Xt} are the sample covariances Both direct and resampling approaches are used to estimate the variance of the sample covariances. This paper compares the performance of these variance estimates. Using a direct approach, we show that a consistent windowed periodogram estimate for the spectrum is more effective than using the periodogram itself. A frequency domain bootstrap for time series is proposed and analyzed, and we introduce a frequency domain version of the jackknife that is shown to be asymptotically unbiased and consistent for Gaussian processes. Monte Carlo techniques show that the time domain jackknife and subseries method cannot be recommended. For a Gaussian underlying series a direct approach using a smoothed periodogram is best; for a non‐Gaussian series the frequency domain bootstrap appears preferable. For small samples, the bootstraps are dangerous: both the direct approach and frequency domain jackknife are better.

Original languageEnglish (US)
Pages (from-to)23-42
Number of pages20
JournalAustralian Journal of Statistics
Volume33
Issue number1
DOIs
StatePublished - 1991

Fingerprint

Autocovariance
Variance Estimation
Resampling
Periodogram
Jackknife
Frequency Domain
Bootstrap
Estimate
Series
Monte Carlo Techniques
Stationary Process
Gaussian Process
Small Sample
Stochastic Processes
Time Domain
Time series

Keywords

  • bootstrap
  • jackknife
  • Sample covariances
  • subseries

ASJC Scopus subject areas

  • Statistics and Probability

Cite this

VARIANCE ESTIMATION FOR SAMPLE AUTOCOVARIANCES : DIRECT AND RESAMPLING APPROACHES. / Hurvich, Clifford M.; Simonoff, Jeffrey S.; Zeger, Scott.

In: Australian Journal of Statistics, Vol. 33, No. 1, 1991, p. 23-42.

Research output: Contribution to journalArticle

Hurvich, Clifford M. ; Simonoff, Jeffrey S. ; Zeger, Scott. / VARIANCE ESTIMATION FOR SAMPLE AUTOCOVARIANCES : DIRECT AND RESAMPLING APPROACHES. In: Australian Journal of Statistics. 1991 ; Vol. 33, No. 1. pp. 23-42.
@article{58658739e7ca44edaccd7f570c4d35d2,
title = "VARIANCE ESTIMATION FOR SAMPLE AUTOCOVARIANCES: DIRECT AND RESAMPLING APPROACHES",
abstract = "The usual covariance estimates for data n‐1 from a stationary zero‐mean stochastic process {Xt} are the sample covariances Both direct and resampling approaches are used to estimate the variance of the sample covariances. This paper compares the performance of these variance estimates. Using a direct approach, we show that a consistent windowed periodogram estimate for the spectrum is more effective than using the periodogram itself. A frequency domain bootstrap for time series is proposed and analyzed, and we introduce a frequency domain version of the jackknife that is shown to be asymptotically unbiased and consistent for Gaussian processes. Monte Carlo techniques show that the time domain jackknife and subseries method cannot be recommended. For a Gaussian underlying series a direct approach using a smoothed periodogram is best; for a non‐Gaussian series the frequency domain bootstrap appears preferable. For small samples, the bootstraps are dangerous: both the direct approach and frequency domain jackknife are better.",
keywords = "bootstrap, jackknife, Sample covariances, subseries",
author = "Hurvich, {Clifford M.} and Simonoff, {Jeffrey S.} and Scott Zeger",
year = "1991",
doi = "10.1111/j.1467-842X.1991.tb00410.x",
language = "English (US)",
volume = "33",
pages = "23--42",
journal = "Australian and New Zealand Journal of Statistics",
issn = "1369-1473",
publisher = "Wiley-Blackwell",
number = "1",

}

TY - JOUR

T1 - VARIANCE ESTIMATION FOR SAMPLE AUTOCOVARIANCES

T2 - DIRECT AND RESAMPLING APPROACHES

AU - Hurvich, Clifford M.

AU - Simonoff, Jeffrey S.

AU - Zeger, Scott

PY - 1991

Y1 - 1991

N2 - The usual covariance estimates for data n‐1 from a stationary zero‐mean stochastic process {Xt} are the sample covariances Both direct and resampling approaches are used to estimate the variance of the sample covariances. This paper compares the performance of these variance estimates. Using a direct approach, we show that a consistent windowed periodogram estimate for the spectrum is more effective than using the periodogram itself. A frequency domain bootstrap for time series is proposed and analyzed, and we introduce a frequency domain version of the jackknife that is shown to be asymptotically unbiased and consistent for Gaussian processes. Monte Carlo techniques show that the time domain jackknife and subseries method cannot be recommended. For a Gaussian underlying series a direct approach using a smoothed periodogram is best; for a non‐Gaussian series the frequency domain bootstrap appears preferable. For small samples, the bootstraps are dangerous: both the direct approach and frequency domain jackknife are better.

AB - The usual covariance estimates for data n‐1 from a stationary zero‐mean stochastic process {Xt} are the sample covariances Both direct and resampling approaches are used to estimate the variance of the sample covariances. This paper compares the performance of these variance estimates. Using a direct approach, we show that a consistent windowed periodogram estimate for the spectrum is more effective than using the periodogram itself. A frequency domain bootstrap for time series is proposed and analyzed, and we introduce a frequency domain version of the jackknife that is shown to be asymptotically unbiased and consistent for Gaussian processes. Monte Carlo techniques show that the time domain jackknife and subseries method cannot be recommended. For a Gaussian underlying series a direct approach using a smoothed periodogram is best; for a non‐Gaussian series the frequency domain bootstrap appears preferable. For small samples, the bootstraps are dangerous: both the direct approach and frequency domain jackknife are better.

KW - bootstrap

KW - jackknife

KW - Sample covariances

KW - subseries

UR - http://www.scopus.com/inward/record.url?scp=84985534870&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84985534870&partnerID=8YFLogxK

U2 - 10.1111/j.1467-842X.1991.tb00410.x

DO - 10.1111/j.1467-842X.1991.tb00410.x

M3 - Article

AN - SCOPUS:84985534870

VL - 33

SP - 23

EP - 42

JO - Australian and New Zealand Journal of Statistics

JF - Australian and New Zealand Journal of Statistics

SN - 1369-1473

IS - 1

ER -