Performance of Bayesian outlier diagnostic in testing mean vector

Rofizah Mohammad, Firdaus Mohamad Hamzah

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The diagnostic measure kd which is used to measure the effect of a single observation d on model choice was applied to a variety of univariate model. The purpose of this study is to assess the performance of this diagnostic measure when applying to multivariate structure for testing the specified mean vector. We illustrate the method using data generated from multivariate normal distribution. If X a p-variate normal random variable of size n with the mean vector θ and a known covariance matrix, we consider the null hypothesis that the mean vector θ is zero. From this simulation we test the performance of kd for several n and p values.

Original languageEnglish
Title of host publicationStatistics and Operational Research International Conference, SORIC 2013
PublisherAmerican Institute of Physics Inc.
Pages275-281
Number of pages7
Volume1613
ISBN (Electronic)9780735412491
DOIs
Publication statusPublished - 1 Jan 2014
EventStatistics and Operational Research International Conference, SORIC 2013 - Sarawak, Malaysia
Duration: 3 Dec 20135 Dec 2013

Other

OtherStatistics and Operational Research International Conference, SORIC 2013
CountryMalaysia
CitySarawak
Period3/12/135/12/13

Fingerprint

null hypothesis
random variables
normal density functions
simulation

Keywords

  • Bayes factor
  • influential observation
  • model choice

ASJC Scopus subject areas

  • Physics and Astronomy(all)

Cite this

Mohammad, R., & Mohamad Hamzah, F. (2014). Performance of Bayesian outlier diagnostic in testing mean vector. In Statistics and Operational Research International Conference, SORIC 2013 (Vol. 1613, pp. 275-281). American Institute of Physics Inc.. https://doi.org/10.1063/1.4894352

Performance of Bayesian outlier diagnostic in testing mean vector. / Mohammad, Rofizah; Mohamad Hamzah, Firdaus.

Statistics and Operational Research International Conference, SORIC 2013. Vol. 1613 American Institute of Physics Inc., 2014. p. 275-281.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Mohammad, R & Mohamad Hamzah, F 2014, Performance of Bayesian outlier diagnostic in testing mean vector. in Statistics and Operational Research International Conference, SORIC 2013. vol. 1613, American Institute of Physics Inc., pp. 275-281, Statistics and Operational Research International Conference, SORIC 2013, Sarawak, Malaysia, 3/12/13. https://doi.org/10.1063/1.4894352
Mohammad R, Mohamad Hamzah F. Performance of Bayesian outlier diagnostic in testing mean vector. In Statistics and Operational Research International Conference, SORIC 2013. Vol. 1613. American Institute of Physics Inc. 2014. p. 275-281 https://doi.org/10.1063/1.4894352
Mohammad, Rofizah ; Mohamad Hamzah, Firdaus. / Performance of Bayesian outlier diagnostic in testing mean vector. Statistics and Operational Research International Conference, SORIC 2013. Vol. 1613 American Institute of Physics Inc., 2014. pp. 275-281
@inproceedings{df7c61c7df3d4b9ba57a0103576ec762,
title = "Performance of Bayesian outlier diagnostic in testing mean vector",
abstract = "The diagnostic measure kd which is used to measure the effect of a single observation d on model choice was applied to a variety of univariate model. The purpose of this study is to assess the performance of this diagnostic measure when applying to multivariate structure for testing the specified mean vector. We illustrate the method using data generated from multivariate normal distribution. If X a p-variate normal random variable of size n with the mean vector θ and a known covariance matrix, we consider the null hypothesis that the mean vector θ is zero. From this simulation we test the performance of kd for several n and p values.",
keywords = "Bayes factor, influential observation, model choice",
author = "Rofizah Mohammad and {Mohamad Hamzah}, Firdaus",
year = "2014",
month = "1",
day = "1",
doi = "10.1063/1.4894352",
language = "English",
volume = "1613",
pages = "275--281",
booktitle = "Statistics and Operational Research International Conference, SORIC 2013",
publisher = "American Institute of Physics Inc.",

}

TY - GEN

T1 - Performance of Bayesian outlier diagnostic in testing mean vector

AU - Mohammad, Rofizah

AU - Mohamad Hamzah, Firdaus

PY - 2014/1/1

Y1 - 2014/1/1

N2 - The diagnostic measure kd which is used to measure the effect of a single observation d on model choice was applied to a variety of univariate model. The purpose of this study is to assess the performance of this diagnostic measure when applying to multivariate structure for testing the specified mean vector. We illustrate the method using data generated from multivariate normal distribution. If X a p-variate normal random variable of size n with the mean vector θ and a known covariance matrix, we consider the null hypothesis that the mean vector θ is zero. From this simulation we test the performance of kd for several n and p values.

AB - The diagnostic measure kd which is used to measure the effect of a single observation d on model choice was applied to a variety of univariate model. The purpose of this study is to assess the performance of this diagnostic measure when applying to multivariate structure for testing the specified mean vector. We illustrate the method using data generated from multivariate normal distribution. If X a p-variate normal random variable of size n with the mean vector θ and a known covariance matrix, we consider the null hypothesis that the mean vector θ is zero. From this simulation we test the performance of kd for several n and p values.

KW - Bayes factor

KW - influential observation

KW - model choice

UR - http://www.scopus.com/inward/record.url?scp=85035745539&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85035745539&partnerID=8YFLogxK

U2 - 10.1063/1.4894352

DO - 10.1063/1.4894352

M3 - Conference contribution

AN - SCOPUS:85035745539

VL - 1613

SP - 275

EP - 281

BT - Statistics and Operational Research International Conference, SORIC 2013

PB - American Institute of Physics Inc.

ER -