Bayesian variable selection in Poisson change-point regression analysis

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

In this article, we develop a Bayesian variable selection method that concerns selection of covariates in the Poisson change-point regression model with both discrete and continuous candidate covariates. Ranging from a null model with no selected covariates to a full model including all covariates, the Bayesian variable selection method searches the entire model space, estimates posterior inclusion probabilities of covariates, and obtains model averaged estimates on coefficients to covariates, while simultaneously estimating a time-varying baseline rate due to change-points. For posterior computation, the Metropolis-Hastings within partially collapsed Gibbs sampler is developed to efficiently fit the Poisson change-point regression model with variable selection. We illustrate the proposed method using simulated and real datasets.

Original languageEnglish
Pages (from-to)2267-2282
Number of pages16
JournalCommunications in Statistics: Simulation and Computation
Volume46
Issue number3
DOIs
Publication statusPublished - 2017 Mar 16

Fingerprint

Change-point Analysis
Bayesian Variable Selection
Regression Analysis
Regression analysis
Covariates
Siméon Denis Poisson
Change-point Model
Regression Model
Inclusion Probabilities
Metropolis-Hastings
Gibbs Sampler
Change Point
Posterior Probability
Variable Selection
Search Methods
Model
Estimate
Null
Baseline
Time-varying

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Modelling and Simulation

Cite this

@article{7890d3108d664a23b48ef57448e10545,
title = "Bayesian variable selection in Poisson change-point regression analysis",
abstract = "In this article, we develop a Bayesian variable selection method that concerns selection of covariates in the Poisson change-point regression model with both discrete and continuous candidate covariates. Ranging from a null model with no selected covariates to a full model including all covariates, the Bayesian variable selection method searches the entire model space, estimates posterior inclusion probabilities of covariates, and obtains model averaged estimates on coefficients to covariates, while simultaneously estimating a time-varying baseline rate due to change-points. For posterior computation, the Metropolis-Hastings within partially collapsed Gibbs sampler is developed to efficiently fit the Poisson change-point regression model with variable selection. We illustrate the proposed method using simulated and real datasets.",
author = "S. Min and Taeyoung Park",
year = "2017",
month = "3",
day = "16",
doi = "10.1080/03610918.2015.1040498",
language = "English",
volume = "46",
pages = "2267--2282",
journal = "Communications in Statistics Part B: Simulation and Computation",
issn = "0361-0918",
publisher = "Taylor and Francis Ltd.",
number = "3",

}

Bayesian variable selection in Poisson change-point regression analysis. / Min, S.; Park, Taeyoung.

In: Communications in Statistics: Simulation and Computation, Vol. 46, No. 3, 16.03.2017, p. 2267-2282.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Bayesian variable selection in Poisson change-point regression analysis

AU - Min, S.

AU - Park, Taeyoung

PY - 2017/3/16

Y1 - 2017/3/16

N2 - In this article, we develop a Bayesian variable selection method that concerns selection of covariates in the Poisson change-point regression model with both discrete and continuous candidate covariates. Ranging from a null model with no selected covariates to a full model including all covariates, the Bayesian variable selection method searches the entire model space, estimates posterior inclusion probabilities of covariates, and obtains model averaged estimates on coefficients to covariates, while simultaneously estimating a time-varying baseline rate due to change-points. For posterior computation, the Metropolis-Hastings within partially collapsed Gibbs sampler is developed to efficiently fit the Poisson change-point regression model with variable selection. We illustrate the proposed method using simulated and real datasets.

AB - In this article, we develop a Bayesian variable selection method that concerns selection of covariates in the Poisson change-point regression model with both discrete and continuous candidate covariates. Ranging from a null model with no selected covariates to a full model including all covariates, the Bayesian variable selection method searches the entire model space, estimates posterior inclusion probabilities of covariates, and obtains model averaged estimates on coefficients to covariates, while simultaneously estimating a time-varying baseline rate due to change-points. For posterior computation, the Metropolis-Hastings within partially collapsed Gibbs sampler is developed to efficiently fit the Poisson change-point regression model with variable selection. We illustrate the proposed method using simulated and real datasets.

UR - http://www.scopus.com/inward/record.url?scp=84996879853&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84996879853&partnerID=8YFLogxK

U2 - 10.1080/03610918.2015.1040498

DO - 10.1080/03610918.2015.1040498

M3 - Article

AN - SCOPUS:84996879853

VL - 46

SP - 2267

EP - 2282

JO - Communications in Statistics Part B: Simulation and Computation

JF - Communications in Statistics Part B: Simulation and Computation

SN - 0361-0918

IS - 3

ER -