Bivariate zero-inflated generalized Poisson regression model with flexible covariance

Pouya Faroughi, Noriszura Ismail

Research output: Contribution to journalArticle

Abstract

This paper introduces several forms of nested bivariate zero-inflated generalized Poisson (BZIGP) regression model which can be fitted to bivariate and zero-inflated count data. The main advantage of having several forms of BZIGP regression model is that they are nested and allow likelihood ratio test to be performed for choosing the best model. In addition, the BZIGP regression models have flexible forms of marginal mean–variance relationship, can be fitted to bivariate and zero-inflated count data with positive or negative correlations, and allow additional overdispersion of the two response variables. The BZIGP regression models are fitted to the Australian Health Survey data.

Original languageEnglish
Pages (from-to)7769-7785
Number of pages17
JournalCommunications in Statistics - Theory and Methods
Volume46
Issue number15
DOIs
Publication statusPublished - 3 Aug 2017

Fingerprint

Poisson Regression
Poisson Model
Regression Model
Zero
Count Data
Overdispersion
Survey Data
Likelihood Ratio Test
Health
Form

Keywords

  • Bivariate counts
  • generalized Poisson
  • health care
  • zero inflation

ASJC Scopus subject areas

  • Statistics and Probability

Cite this

Bivariate zero-inflated generalized Poisson regression model with flexible covariance. / Faroughi, Pouya; Ismail, Noriszura.

In: Communications in Statistics - Theory and Methods, Vol. 46, No. 15, 03.08.2017, p. 7769-7785.

Research output: Contribution to journalArticle

@article{6d9daee004cb46659107465aa8255565,
title = "Bivariate zero-inflated generalized Poisson regression model with flexible covariance",
abstract = "This paper introduces several forms of nested bivariate zero-inflated generalized Poisson (BZIGP) regression model which can be fitted to bivariate and zero-inflated count data. The main advantage of having several forms of BZIGP regression model is that they are nested and allow likelihood ratio test to be performed for choosing the best model. In addition, the BZIGP regression models have flexible forms of marginal mean–variance relationship, can be fitted to bivariate and zero-inflated count data with positive or negative correlations, and allow additional overdispersion of the two response variables. The BZIGP regression models are fitted to the Australian Health Survey data.",
keywords = "Bivariate counts, generalized Poisson, health care, zero inflation",
author = "Pouya Faroughi and Noriszura Ismail",
year = "2017",
month = "8",
day = "3",
doi = "10.1080/03610926.2016.1165846",
language = "English",
volume = "46",
pages = "7769--7785",
journal = "Communications in Statistics - Theory and Methods",
issn = "0361-0926",
publisher = "Taylor and Francis Ltd.",
number = "15",

}

TY - JOUR

T1 - Bivariate zero-inflated generalized Poisson regression model with flexible covariance

AU - Faroughi, Pouya

AU - Ismail, Noriszura

PY - 2017/8/3

Y1 - 2017/8/3

N2 - This paper introduces several forms of nested bivariate zero-inflated generalized Poisson (BZIGP) regression model which can be fitted to bivariate and zero-inflated count data. The main advantage of having several forms of BZIGP regression model is that they are nested and allow likelihood ratio test to be performed for choosing the best model. In addition, the BZIGP regression models have flexible forms of marginal mean–variance relationship, can be fitted to bivariate and zero-inflated count data with positive or negative correlations, and allow additional overdispersion of the two response variables. The BZIGP regression models are fitted to the Australian Health Survey data.

AB - This paper introduces several forms of nested bivariate zero-inflated generalized Poisson (BZIGP) regression model which can be fitted to bivariate and zero-inflated count data. The main advantage of having several forms of BZIGP regression model is that they are nested and allow likelihood ratio test to be performed for choosing the best model. In addition, the BZIGP regression models have flexible forms of marginal mean–variance relationship, can be fitted to bivariate and zero-inflated count data with positive or negative correlations, and allow additional overdispersion of the two response variables. The BZIGP regression models are fitted to the Australian Health Survey data.

KW - Bivariate counts

KW - generalized Poisson

KW - health care

KW - zero inflation

UR - http://www.scopus.com/inward/record.url?scp=85018850469&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85018850469&partnerID=8YFLogxK

U2 - 10.1080/03610926.2016.1165846

DO - 10.1080/03610926.2016.1165846

M3 - Article

AN - SCOPUS:85018850469

VL - 46

SP - 7769

EP - 7785

JO - Communications in Statistics - Theory and Methods

JF - Communications in Statistics - Theory and Methods

SN - 0361-0926

IS - 15

ER -