Master of Science (MS)
First Committee Member
Ashok K. Singh
Number of Pages
Extensive work has been done on goodness-of-fit (GOF) tests for data assumed to have come from univariate continuous distributions; however, literature on GOF procedures for univariate discrete distributions is rather sparse in compariSon The Poisson distribution in particular has received much attention in the study of GOF tests due to its numerous applications as a model for observable phenomena. Hence, we survey existing GOF tests for Poissonity and present a useful guide to the most commonly used distribution-free GOF tests in practice. We then propose and investigate a graphical test of fit for the Poisson model that is based on a Poisson Q-Q plot, a squared correlation coefficient R2 test statistic, and a sampling distribution of the R2 test statistic simulated by parametric bootstrap. Similar methods exist for continuous distributions like the univariate normal and extreme-value distributions under regression tests of fit. Simulated examples as well as historically well-known Poisson data sets are then used to illustrate the proposed goodness-of-fit test for Poissonity.
Approach; Fit; Goodness; Graphical; Model; Poisson
University of Nevada, Las Vegas
If you are the rightful copyright holder of this dissertation or thesis and wish to have the full text removed from Digital Scholarship@UNLV, please submit a request to firstname.lastname@example.org and include clear identification of the work, preferably with URL.
Padilla, Davin P, "A graphical approach for goodness-of-fit of Poisson model" (2003). UNLV Retrospective Theses & Dissertations. 1607.