A new look at portmanteau tests

Fumiya Akashi, Hiroaki Odashima, Masanobu Taniguchi, Anna Clara Monti

Research output: Contribution to journalArticle

Abstract

Portmanteau tests are some of the most commonly used statistical methods for model diagnostics. They can be applied in model checking either in the time series or in the regression context. The present paper proposes a portmanteau-type test, based on a sort of likelihood ratio statistic, useful to test general parametric hypotheses inherent to statistical models, which includes the classical portmanteau tests as special cases. Sufficient conditions for the statistic to be asymptotically chi-square distributed are elucidated in terms of the Fisher information matrix, and the results have very clear implications for the relationships between the parameter of interest and nuisance parameter. In addition, the power of the test is investigated when local alternative hypotheses are considered. Some interesting applications of the proposed test to various problems are illustrated, such as serial correlation tests where the proposed test is shown to be asymptotically equivalent to classical tests. Since portmanteau tests are widely used in many fields, it appears essential to elucidate the fundamental mechanism in a unified view.

Original languageEnglish
Pages (from-to)121-137
Number of pages17
JournalSankhya A
Volume80
Issue number1
DOIs
Publication statusPublished - 2018 Jan 1

Fingerprint

Portmanteau Test
Model Diagnostics
Local Alternatives
Serial Correlation
Fisher Information Matrix
Likelihood Ratio Statistic
Asymptotically equivalent
Chi-square
Nuisance Parameter
Portmanteau test
Statistical method
Sort
Model Checking
Statistical Model
Statistic
Time series
Regression
Sufficient Conditions

Keywords

  • Asymptotic local power
  • Portmanteau test
  • Serial correlation
  • Time series analysis
  • Variable selection

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

Akashi, F., Odashima, H., Taniguchi, M., & Monti, A. C. (2018). A new look at portmanteau tests. Sankhya A, 80(1), 121-137. https://doi.org/10.1007/s13171-017-0109-3

A new look at portmanteau tests. / Akashi, Fumiya; Odashima, Hiroaki; Taniguchi, Masanobu; Monti, Anna Clara.

In: Sankhya A, Vol. 80, No. 1, 01.01.2018, p. 121-137.

Research output: Contribution to journalArticle

Akashi, F, Odashima, H, Taniguchi, M & Monti, AC 2018, 'A new look at portmanteau tests', Sankhya A, vol. 80, no. 1, pp. 121-137. https://doi.org/10.1007/s13171-017-0109-3
Akashi, Fumiya ; Odashima, Hiroaki ; Taniguchi, Masanobu ; Monti, Anna Clara. / A new look at portmanteau tests. In: Sankhya A. 2018 ; Vol. 80, No. 1. pp. 121-137.
@article{da0da57a3b9941a482d684c8053cb9e8,
title = "A new look at portmanteau tests",
abstract = "Portmanteau tests are some of the most commonly used statistical methods for model diagnostics. They can be applied in model checking either in the time series or in the regression context. The present paper proposes a portmanteau-type test, based on a sort of likelihood ratio statistic, useful to test general parametric hypotheses inherent to statistical models, which includes the classical portmanteau tests as special cases. Sufficient conditions for the statistic to be asymptotically chi-square distributed are elucidated in terms of the Fisher information matrix, and the results have very clear implications for the relationships between the parameter of interest and nuisance parameter. In addition, the power of the test is investigated when local alternative hypotheses are considered. Some interesting applications of the proposed test to various problems are illustrated, such as serial correlation tests where the proposed test is shown to be asymptotically equivalent to classical tests. Since portmanteau tests are widely used in many fields, it appears essential to elucidate the fundamental mechanism in a unified view.",
keywords = "Asymptotic local power, Portmanteau test, Serial correlation, Time series analysis, Variable selection",
author = "Fumiya Akashi and Hiroaki Odashima and Masanobu Taniguchi and Monti, {Anna Clara}",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/s13171-017-0109-3",
language = "English",
volume = "80",
pages = "121--137",
journal = "Sankhya A",
issn = "0976-836X",
publisher = "Springer India",
number = "1",

}

TY - JOUR

T1 - A new look at portmanteau tests

AU - Akashi, Fumiya

AU - Odashima, Hiroaki

AU - Taniguchi, Masanobu

AU - Monti, Anna Clara

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Portmanteau tests are some of the most commonly used statistical methods for model diagnostics. They can be applied in model checking either in the time series or in the regression context. The present paper proposes a portmanteau-type test, based on a sort of likelihood ratio statistic, useful to test general parametric hypotheses inherent to statistical models, which includes the classical portmanteau tests as special cases. Sufficient conditions for the statistic to be asymptotically chi-square distributed are elucidated in terms of the Fisher information matrix, and the results have very clear implications for the relationships between the parameter of interest and nuisance parameter. In addition, the power of the test is investigated when local alternative hypotheses are considered. Some interesting applications of the proposed test to various problems are illustrated, such as serial correlation tests where the proposed test is shown to be asymptotically equivalent to classical tests. Since portmanteau tests are widely used in many fields, it appears essential to elucidate the fundamental mechanism in a unified view.

AB - Portmanteau tests are some of the most commonly used statistical methods for model diagnostics. They can be applied in model checking either in the time series or in the regression context. The present paper proposes a portmanteau-type test, based on a sort of likelihood ratio statistic, useful to test general parametric hypotheses inherent to statistical models, which includes the classical portmanteau tests as special cases. Sufficient conditions for the statistic to be asymptotically chi-square distributed are elucidated in terms of the Fisher information matrix, and the results have very clear implications for the relationships between the parameter of interest and nuisance parameter. In addition, the power of the test is investigated when local alternative hypotheses are considered. Some interesting applications of the proposed test to various problems are illustrated, such as serial correlation tests where the proposed test is shown to be asymptotically equivalent to classical tests. Since portmanteau tests are widely used in many fields, it appears essential to elucidate the fundamental mechanism in a unified view.

KW - Asymptotic local power

KW - Portmanteau test

KW - Serial correlation

KW - Time series analysis

KW - Variable selection

UR - http://www.scopus.com/inward/record.url?scp=85053552890&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85053552890&partnerID=8YFLogxK

U2 - 10.1007/s13171-017-0109-3

DO - 10.1007/s13171-017-0109-3

M3 - Article

AN - SCOPUS:85034051360

VL - 80

SP - 121

EP - 137

JO - Sankhya A

JF - Sankhya A

SN - 0976-836X

IS - 1

ER -