Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


16227.pdf (1015.31 KB)

ETD Abstract Container

Abstract Header

Bayesian Model Selection for Poisson and Related Models

Guo, Yixuan
http://rave.ohiolink.edu/etdc/view?acc_num=ucin1439310177

Abstract Details

2015, PhD, University of Cincinnati, Arts and Sciences: Mathematical Sciences.
Poisson model is one of the fundamental discrete models used to model count data in various fields. It assumes that the mean and variance of data are approximately equal. In practice, the observed data often violates this assumption because the variance can be larger than the mean commonly referred to as over-dispersion. Several models have been developed based on the Poisson model to address the issue of dispersion occurred in data; generalized Poisson model, zero-inflated Poisson model, and zero-inflated generalized Poisson model are examples of such distributions. In this thesis, I will focus on developing a method within the Bayesian framework to compare these four models. This method is generic and can be readily generalized to the comparison of any number of models. We will use non-informative prior and importance sampling to calculate the posterior probability for each model. We also use the same method to compare regression models, namely Poisson, generalized Poisson and negative binomial regression models.
Siva Sivaganesan, Ph.D. (Committee Chair)
Emily Kang, Ph.D. (Committee Member)
Seongho Song, Ph.D. (Committee Member)
186 p.
Statistics
Model selection; Posterior probability; Jeffreys prior; Generalized Poisson model; Zero-inflated Poisson model; Zero-inflated generalized Poisson model

Recommended Citations

Citations

  • Guo, Y. (2015). Bayesian Model Selection for Poisson and Related Models [Doctoral dissertation, University of Cincinnati]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1439310177

    APA Style (7th edition)

  • Guo, Yixuan. Bayesian Model Selection for Poisson and Related Models. 2015. University of Cincinnati, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=ucin1439310177.

    MLA Style (8th edition)

  • Guo, Yixuan. "Bayesian Model Selection for Poisson and Related Models." Doctoral dissertation, University of Cincinnati, 2015. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1439310177

    Chicago Manual of Style (17th edition)

ucin1439310177
7,972
© 2015, all rights reserved.
This open access ETD is published by University of Cincinnati and OhioLINK.