Tests for Scale Changes Based on Pairwise Differences

Carina Gerstenberger, Daniel Vogel (Corresponding Author), Martin Wendler

Research output: Contribution to journalArticle

Abstract

In many applications it is important to know whether the amount of fluctuation in a series of observations changes over time. In this article, we investigate different tests for detecting changes in the scale of mean-stationary time series. The classical approach, based on the CUSUM test applied to the squared centered observations, is very vulnerable to outliers and impractical for heavy-tailed data, which leads us to contemplate test statistics based on alternative, less outlier-sensitive scale estimators. It turns out that the tests based on Gini’s mean difference (the average of all pairwise distances) and generalized Qn estimators (sample quantiles of all pairwise distances) are very suitable candidates. They improve upon the classical test not only under heavy tails or in the presence of outliers, but also under normality.

We use recent results on the process convergence of U-statistics and U-quantiles for dependent sequences to derive the limiting distribution of the test statistics and propose estimators for the long-run variance. We show the consistency of the tests and demonstrate the applicability of the new change-point detection methods at two real-life data examples from hydrology and finance.
Original languageEnglish
JournalJournal of the American Statistical Association
Early online date21 Aug 2019
DOIs
Publication statusE-pub ahead of print - 21 Aug 2019

Fingerprint

Pairwise
Outlier
Estimator
Test Statistic
CUSUM Test
Sample Quantiles
Change-point Detection
Stationary Time Series
Hydrology
Heavy Tails
U-statistics
Quantile
Long-run
Limiting Distribution
Finance
Normality
Fluctuations
Series
Dependent
Alternatives

Keywords

  • Asymptotic relative efficiency
  • Change-point analysis
  • Gini’s mean difference
  • Long-run variance estimation
  • Median absolute deviation
  • Qn scale estimator
  • U-quantile
  • U-statistic
  • Block bootstrap
  • Gini's mean difference
  • QUANTILES
  • SUMS
  • BOOTSTRAP
  • ESTIMATORS
  • VARIANCE
  • SQUARES

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

Tests for Scale Changes Based on Pairwise Differences. / Gerstenberger, Carina; Vogel, Daniel (Corresponding Author); Wendler, Martin.

In: Journal of the American Statistical Association, 21.08.2019.

Research output: Contribution to journalArticle

@article{a9b8d022ee3648d1a01b28ad3686654d,
title = "Tests for Scale Changes Based on Pairwise Differences",
abstract = "In many applications it is important to know whether the amount of fluctuation in a series of observations changes over time. In this article, we investigate different tests for detecting changes in the scale of mean-stationary time series. The classical approach, based on the CUSUM test applied to the squared centered observations, is very vulnerable to outliers and impractical for heavy-tailed data, which leads us to contemplate test statistics based on alternative, less outlier-sensitive scale estimators. It turns out that the tests based on Gini’s mean difference (the average of all pairwise distances) and generalized Qn estimators (sample quantiles of all pairwise distances) are very suitable candidates. They improve upon the classical test not only under heavy tails or in the presence of outliers, but also under normality.We use recent results on the process convergence of U-statistics and U-quantiles for dependent sequences to derive the limiting distribution of the test statistics and propose estimators for the long-run variance. We show the consistency of the tests and demonstrate the applicability of the new change-point detection methods at two real-life data examples from hydrology and finance.",
keywords = "Asymptotic relative efficiency, Change-point analysis, Gini’s mean difference, Long-run variance estimation, Median absolute deviation, Qn scale estimator, U-quantile, U-statistic, Block bootstrap, Gini's mean difference, QUANTILES, SUMS, BOOTSTRAP, ESTIMATORS, VARIANCE, SQUARES",
author = "Carina Gerstenberger and Daniel Vogel and Martin Wendler",
note = "Acknowledgement The researcher were supported by the Collaborative Research Centre 823 Statistical modelling of nonlinear dynamic processes and the Konrad-Adenauer-Stiftung. The authors thank Svenja Fischer for providing the river Rhine discharge data set and Marco Thiel for the stock exchange data set.",
year = "2019",
month = "8",
day = "21",
doi = "10.1080/01621459.2019.1629938",
language = "English",
journal = "Journal of the American Statistical Association",
issn = "0162-1459",
publisher = "Taylor and Francis Ltd.",

}

TY - JOUR

T1 - Tests for Scale Changes Based on Pairwise Differences

AU - Gerstenberger, Carina

AU - Vogel, Daniel

AU - Wendler, Martin

N1 - Acknowledgement The researcher were supported by the Collaborative Research Centre 823 Statistical modelling of nonlinear dynamic processes and the Konrad-Adenauer-Stiftung. The authors thank Svenja Fischer for providing the river Rhine discharge data set and Marco Thiel for the stock exchange data set.

PY - 2019/8/21

Y1 - 2019/8/21

N2 - In many applications it is important to know whether the amount of fluctuation in a series of observations changes over time. In this article, we investigate different tests for detecting changes in the scale of mean-stationary time series. The classical approach, based on the CUSUM test applied to the squared centered observations, is very vulnerable to outliers and impractical for heavy-tailed data, which leads us to contemplate test statistics based on alternative, less outlier-sensitive scale estimators. It turns out that the tests based on Gini’s mean difference (the average of all pairwise distances) and generalized Qn estimators (sample quantiles of all pairwise distances) are very suitable candidates. They improve upon the classical test not only under heavy tails or in the presence of outliers, but also under normality.We use recent results on the process convergence of U-statistics and U-quantiles for dependent sequences to derive the limiting distribution of the test statistics and propose estimators for the long-run variance. We show the consistency of the tests and demonstrate the applicability of the new change-point detection methods at two real-life data examples from hydrology and finance.

AB - In many applications it is important to know whether the amount of fluctuation in a series of observations changes over time. In this article, we investigate different tests for detecting changes in the scale of mean-stationary time series. The classical approach, based on the CUSUM test applied to the squared centered observations, is very vulnerable to outliers and impractical for heavy-tailed data, which leads us to contemplate test statistics based on alternative, less outlier-sensitive scale estimators. It turns out that the tests based on Gini’s mean difference (the average of all pairwise distances) and generalized Qn estimators (sample quantiles of all pairwise distances) are very suitable candidates. They improve upon the classical test not only under heavy tails or in the presence of outliers, but also under normality.We use recent results on the process convergence of U-statistics and U-quantiles for dependent sequences to derive the limiting distribution of the test statistics and propose estimators for the long-run variance. We show the consistency of the tests and demonstrate the applicability of the new change-point detection methods at two real-life data examples from hydrology and finance.

KW - Asymptotic relative efficiency

KW - Change-point analysis

KW - Gini’s mean difference

KW - Long-run variance estimation

KW - Median absolute deviation

KW - Qn scale estimator

KW - U-quantile

KW - U-statistic

KW - Block bootstrap

KW - Gini's mean difference

KW - QUANTILES

KW - SUMS

KW - BOOTSTRAP

KW - ESTIMATORS

KW - VARIANCE

KW - SQUARES

UR - http://www.mendeley.com/research/tests-scale-changes-based-pairwise-differences

UR - http://www.scopus.com/inward/record.url?scp=85071018762&partnerID=8YFLogxK

U2 - 10.1080/01621459.2019.1629938

DO - 10.1080/01621459.2019.1629938

M3 - Article

JO - Journal of the American Statistical Association

JF - Journal of the American Statistical Association

SN - 0162-1459

ER -