Estimation of count data using mixed Poisson, generalized Poisson and finite Poisson mixture regression models

Hossein Zamani, Pouya Faroughi, Noriszura Ismail

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

1 Citation (Scopus)

Abstract

This study relates the Poisson, mixed Poisson (MP), generalized Poisson (GP) and finite Poisson mixture (FPM) regression models through mean-variance relationship, and suggests the application of these models for overdispersed count data. As an illustration, the regression models are fitted to the US skin care count data. The results indicate that FPM regression model is the best model since it provides the largest log likelihood and the smallest AIC, followed by Poisson-Inverse Gaussion (PIG), GP and negative binomial (NB) regression models. The results also show that NB, PIG and GP regression models provide similar results.

Original languageEnglish
Title of host publicationAIP Conference Proceedings
PublisherAmerican Institute of Physics Inc.
Pages1144-1150
Number of pages7
Volume1602
ISBN (Print)9780735412361
DOIs
Publication statusPublished - 2014
Event3rd International Conference on Mathematical Sciences, ICMS 2013 - Kuala Lumpur
Duration: 17 Dec 201319 Dec 2013

Other

Other3rd International Conference on Mathematical Sciences, ICMS 2013
CityKuala Lumpur
Period17/12/1319/12/13

Fingerprint

regression analysis

Keywords

  • finite Poisson mixture
  • generalized Poisson
  • mixed Poisson
  • Poisson
  • regression model

ASJC Scopus subject areas

  • Physics and Astronomy(all)

Cite this

Zamani, H., Faroughi, P., & Ismail, N. (2014). Estimation of count data using mixed Poisson, generalized Poisson and finite Poisson mixture regression models. In AIP Conference Proceedings (Vol. 1602, pp. 1144-1150). American Institute of Physics Inc.. https://doi.org/10.1063/1.4882628

Estimation of count data using mixed Poisson, generalized Poisson and finite Poisson mixture regression models. / Zamani, Hossein; Faroughi, Pouya; Ismail, Noriszura.

AIP Conference Proceedings. Vol. 1602 American Institute of Physics Inc., 2014. p. 1144-1150.

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

Zamani, H, Faroughi, P & Ismail, N 2014, Estimation of count data using mixed Poisson, generalized Poisson and finite Poisson mixture regression models. in AIP Conference Proceedings. vol. 1602, American Institute of Physics Inc., pp. 1144-1150, 3rd International Conference on Mathematical Sciences, ICMS 2013, Kuala Lumpur, 17/12/13. https://doi.org/10.1063/1.4882628
Zamani H, Faroughi P, Ismail N. Estimation of count data using mixed Poisson, generalized Poisson and finite Poisson mixture regression models. In AIP Conference Proceedings. Vol. 1602. American Institute of Physics Inc. 2014. p. 1144-1150 https://doi.org/10.1063/1.4882628
Zamani, Hossein ; Faroughi, Pouya ; Ismail, Noriszura. / Estimation of count data using mixed Poisson, generalized Poisson and finite Poisson mixture regression models. AIP Conference Proceedings. Vol. 1602 American Institute of Physics Inc., 2014. pp. 1144-1150
@inproceedings{eacbc7825bdd4a4bb25600513285ee57,
title = "Estimation of count data using mixed Poisson, generalized Poisson and finite Poisson mixture regression models",
abstract = "This study relates the Poisson, mixed Poisson (MP), generalized Poisson (GP) and finite Poisson mixture (FPM) regression models through mean-variance relationship, and suggests the application of these models for overdispersed count data. As an illustration, the regression models are fitted to the US skin care count data. The results indicate that FPM regression model is the best model since it provides the largest log likelihood and the smallest AIC, followed by Poisson-Inverse Gaussion (PIG), GP and negative binomial (NB) regression models. The results also show that NB, PIG and GP regression models provide similar results.",
keywords = "finite Poisson mixture, generalized Poisson, mixed Poisson, Poisson, regression model",
author = "Hossein Zamani and Pouya Faroughi and Noriszura Ismail",
year = "2014",
doi = "10.1063/1.4882628",
language = "English",
isbn = "9780735412361",
volume = "1602",
pages = "1144--1150",
booktitle = "AIP Conference Proceedings",
publisher = "American Institute of Physics Inc.",

}

TY - GEN

T1 - Estimation of count data using mixed Poisson, generalized Poisson and finite Poisson mixture regression models

AU - Zamani, Hossein

AU - Faroughi, Pouya

AU - Ismail, Noriszura

PY - 2014

Y1 - 2014

N2 - This study relates the Poisson, mixed Poisson (MP), generalized Poisson (GP) and finite Poisson mixture (FPM) regression models through mean-variance relationship, and suggests the application of these models for overdispersed count data. As an illustration, the regression models are fitted to the US skin care count data. The results indicate that FPM regression model is the best model since it provides the largest log likelihood and the smallest AIC, followed by Poisson-Inverse Gaussion (PIG), GP and negative binomial (NB) regression models. The results also show that NB, PIG and GP regression models provide similar results.

AB - This study relates the Poisson, mixed Poisson (MP), generalized Poisson (GP) and finite Poisson mixture (FPM) regression models through mean-variance relationship, and suggests the application of these models for overdispersed count data. As an illustration, the regression models are fitted to the US skin care count data. The results indicate that FPM regression model is the best model since it provides the largest log likelihood and the smallest AIC, followed by Poisson-Inverse Gaussion (PIG), GP and negative binomial (NB) regression models. The results also show that NB, PIG and GP regression models provide similar results.

KW - finite Poisson mixture

KW - generalized Poisson

KW - mixed Poisson

KW - Poisson

KW - regression model

UR - http://www.scopus.com/inward/record.url?scp=84904093097&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84904093097&partnerID=8YFLogxK

U2 - 10.1063/1.4882628

DO - 10.1063/1.4882628

M3 - Conference contribution

SN - 9780735412361

VL - 1602

SP - 1144

EP - 1150

BT - AIP Conference Proceedings

PB - American Institute of Physics Inc.

ER -