Testing for central symmetry and inference of the unknown center

Xinjie Dai, Cuizhen Niu, Xu Guo

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

In this paper, we consider testing for central symmetry and inference of the unknown center with multivariate data. Our proposed test statistics are based on weighted integrals of empirical characteristic functions. With two special weight functions, we obtain test statistics with simple and closed forms. The test statistics are easy to implement. In fact, they are based merely on pairwise distances between points in the sample. The asymptotic results are developed. It is proven that our proposed tests can converge to finite limit at the rate of n−1 under the null hypothesis and can detect any fixed alternatives. For the unknown center, we also propose two classes of minimum distance estimators based on the previously introduced test statistics. The asymptotic normalities are derived. Efficient algorithms are also developed to compute the estimators in practice. We further consider checking whether the unknown center is equal to a specified value μ0. Extensive simulation studies and one medical data analysis are conducted to illustrate the merits of the proposed methods.

Original languageEnglish (US)
Pages (from-to)15-31
Number of pages17
JournalComputational Statistics and Data Analysis
Volume127
DOIs
StatePublished - Nov 2018

Fingerprint

Test Statistic
Statistics
Symmetry
Unknown
Testing
Minimum Distance Estimator
Empirical Characteristic Function
Multivariate Data
Special Functions
Asymptotic Normality
Null hypothesis
Weight Function
Pairwise
Data analysis
Closed-form
Efficient Algorithms
Simulation Study
Converge
Estimator
Alternatives

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Computational Mathematics
  • Computational Theory and Mathematics
  • Applied Mathematics

Cite this

@article{ed26809319f54a8c9d3c67a8f98e0cf3,
title = "Testing for central symmetry and inference of the unknown center",
abstract = "In this paper, we consider testing for central symmetry and inference of the unknown center with multivariate data. Our proposed test statistics are based on weighted integrals of empirical characteristic functions. With two special weight functions, we obtain test statistics with simple and closed forms. The test statistics are easy to implement. In fact, they are based merely on pairwise distances between points in the sample. The asymptotic results are developed. It is proven that our proposed tests can converge to finite limit at the rate of n−1 under the null hypothesis and can detect any fixed alternatives. For the unknown center, we also propose two classes of minimum distance estimators based on the previously introduced test statistics. The asymptotic normalities are derived. Efficient algorithms are also developed to compute the estimators in practice. We further consider checking whether the unknown center is equal to a specified value μ0. Extensive simulation studies and one medical data analysis are conducted to illustrate the merits of the proposed methods.",
author = "Xinjie Dai and Cuizhen Niu and Xu Guo",
year = "2018",
month = "11",
doi = "10.1016/j.csda.2018.05.007",
language = "English (US)",
volume = "127",
pages = "15--31",
journal = "Computational Statistics and Data Analysis",
issn = "0167-9473",
publisher = "Elsevier",

}

Testing for central symmetry and inference of the unknown center. / Dai, Xinjie; Niu, Cuizhen; Guo, Xu.

In: Computational Statistics and Data Analysis, Vol. 127, 11.2018, p. 15-31.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Testing for central symmetry and inference of the unknown center

AU - Dai, Xinjie

AU - Niu, Cuizhen

AU - Guo, Xu

PY - 2018/11

Y1 - 2018/11

N2 - In this paper, we consider testing for central symmetry and inference of the unknown center with multivariate data. Our proposed test statistics are based on weighted integrals of empirical characteristic functions. With two special weight functions, we obtain test statistics with simple and closed forms. The test statistics are easy to implement. In fact, they are based merely on pairwise distances between points in the sample. The asymptotic results are developed. It is proven that our proposed tests can converge to finite limit at the rate of n−1 under the null hypothesis and can detect any fixed alternatives. For the unknown center, we also propose two classes of minimum distance estimators based on the previously introduced test statistics. The asymptotic normalities are derived. Efficient algorithms are also developed to compute the estimators in practice. We further consider checking whether the unknown center is equal to a specified value μ0. Extensive simulation studies and one medical data analysis are conducted to illustrate the merits of the proposed methods.

AB - In this paper, we consider testing for central symmetry and inference of the unknown center with multivariate data. Our proposed test statistics are based on weighted integrals of empirical characteristic functions. With two special weight functions, we obtain test statistics with simple and closed forms. The test statistics are easy to implement. In fact, they are based merely on pairwise distances between points in the sample. The asymptotic results are developed. It is proven that our proposed tests can converge to finite limit at the rate of n−1 under the null hypothesis and can detect any fixed alternatives. For the unknown center, we also propose two classes of minimum distance estimators based on the previously introduced test statistics. The asymptotic normalities are derived. Efficient algorithms are also developed to compute the estimators in practice. We further consider checking whether the unknown center is equal to a specified value μ0. Extensive simulation studies and one medical data analysis are conducted to illustrate the merits of the proposed methods.

UR - http://www.scopus.com/inward/record.url?scp=85047599002&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85047599002&partnerID=8YFLogxK

U2 - 10.1016/j.csda.2018.05.007

DO - 10.1016/j.csda.2018.05.007

M3 - Article

AN - SCOPUS:85047599002

VL - 127

SP - 15

EP - 31

JO - Computational Statistics and Data Analysis

JF - Computational Statistics and Data Analysis

SN - 0167-9473

ER -