Inference for Multivariate Regression Model based on Synthetic Data generated under Fixed-Posterior Predictive Sampling

Comparison with Plug-in Sampling

Ricardo Moura, Klein Martin, Bimal Sinha, Carlos A. Coelho

Research output: Contribution to journalArticle

13 Downloads (Pure)

Abstract

The authors derive likelihood-based exact inference methods for the multivariate regression model, for singly imputed synthetic data generated via Posterior Predictive Sampling (PPS) and for multiply imputed synthetic data generated via a newly proposed sampling method, which the authors call Fixed-Posterior Predictive Sampling (FPPS). In the single imputation case, our proposed FPPS method concurs with the usual Posterior Predictive Sampling (PPS) method, thus filling the gap in the existing literature where inferential methods are only available for multiple imputation. Simulation studies compare the results obtained with those for the exact test procedures under the Plug-in Sampling method, obtained by the same authors. Measures of privacy are discussed and compared with the measures derived for the Plug-in Sampling method. An application using U.S. 2000 Current Population Survey data is discussed.
Original languageEnglish
Pages (from-to)155-186
Number of pages32
JournalREVSTAT - STATISTICAL JOURNAL
Volume15
Issue number2
Publication statusPublished - 1 Apr 2017

Fingerprint

sampling
method
comparison
simulation

Keywords

  • finite sample inference
  • maximum likelihood estimation
  • pivotal quantity
  • plug-in sampling
  • statistical disclosure control
  • unbiased estimators

Cite this

@article{b71f594242b94dbfb9a177dbacc4ce62,
title = "Inference for Multivariate Regression Model based on Synthetic Data generated under Fixed-Posterior Predictive Sampling: Comparison with Plug-in Sampling",
abstract = "The authors derive likelihood-based exact inference methods for the multivariate regression model, for singly imputed synthetic data generated via Posterior Predictive Sampling (PPS) and for multiply imputed synthetic data generated via a newly proposed sampling method, which the authors call Fixed-Posterior Predictive Sampling (FPPS). In the single imputation case, our proposed FPPS method concurs with the usual Posterior Predictive Sampling (PPS) method, thus filling the gap in the existing literature where inferential methods are only available for multiple imputation. Simulation studies compare the results obtained with those for the exact test procedures under the Plug-in Sampling method, obtained by the same authors. Measures of privacy are discussed and compared with the measures derived for the Plug-in Sampling method. An application using U.S. 2000 Current Population Survey data is discussed.",
keywords = "finite sample inference, maximum likelihood estimation, pivotal quantity, plug-in sampling, statistical disclosure control, unbiased estimators",
author = "Ricardo Moura and Klein Martin and Bimal Sinha and Coelho, {Carlos A.}",
note = "Portuguese Foundation for Science and Technology - UID/MAT/00297/2013",
year = "2017",
month = "4",
day = "1",
language = "English",
volume = "15",
pages = "155--186",
journal = "REVSTAT - STATISTICAL JOURNAL",
issn = "1645-6726",
publisher = "INST NACIONAL ESTATISTICA-INE",
number = "2",

}

Inference for Multivariate Regression Model based on Synthetic Data generated under Fixed-Posterior Predictive Sampling : Comparison with Plug-in Sampling. / Moura, Ricardo; Martin, Klein; Sinha, Bimal; Coelho, Carlos A.

In: REVSTAT - STATISTICAL JOURNAL, Vol. 15, No. 2, 01.04.2017, p. 155-186.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Inference for Multivariate Regression Model based on Synthetic Data generated under Fixed-Posterior Predictive Sampling

T2 - Comparison with Plug-in Sampling

AU - Moura, Ricardo

AU - Martin, Klein

AU - Sinha, Bimal

AU - Coelho, Carlos A.

N1 - Portuguese Foundation for Science and Technology - UID/MAT/00297/2013

PY - 2017/4/1

Y1 - 2017/4/1

N2 - The authors derive likelihood-based exact inference methods for the multivariate regression model, for singly imputed synthetic data generated via Posterior Predictive Sampling (PPS) and for multiply imputed synthetic data generated via a newly proposed sampling method, which the authors call Fixed-Posterior Predictive Sampling (FPPS). In the single imputation case, our proposed FPPS method concurs with the usual Posterior Predictive Sampling (PPS) method, thus filling the gap in the existing literature where inferential methods are only available for multiple imputation. Simulation studies compare the results obtained with those for the exact test procedures under the Plug-in Sampling method, obtained by the same authors. Measures of privacy are discussed and compared with the measures derived for the Plug-in Sampling method. An application using U.S. 2000 Current Population Survey data is discussed.

AB - The authors derive likelihood-based exact inference methods for the multivariate regression model, for singly imputed synthetic data generated via Posterior Predictive Sampling (PPS) and for multiply imputed synthetic data generated via a newly proposed sampling method, which the authors call Fixed-Posterior Predictive Sampling (FPPS). In the single imputation case, our proposed FPPS method concurs with the usual Posterior Predictive Sampling (PPS) method, thus filling the gap in the existing literature where inferential methods are only available for multiple imputation. Simulation studies compare the results obtained with those for the exact test procedures under the Plug-in Sampling method, obtained by the same authors. Measures of privacy are discussed and compared with the measures derived for the Plug-in Sampling method. An application using U.S. 2000 Current Population Survey data is discussed.

KW - finite sample inference

KW - maximum likelihood estimation

KW - pivotal quantity

KW - plug-in sampling

KW - statistical disclosure control

KW - unbiased estimators

M3 - Article

VL - 15

SP - 155

EP - 186

JO - REVSTAT - STATISTICAL JOURNAL

JF - REVSTAT - STATISTICAL JOURNAL

SN - 1645-6726

IS - 2

ER -