Abstract
THESIS: Developing A Testing Procedure for the Parameters of Bivariate
Geometric Distribution
STUDENT: Md. Fitrat Hossain
DEGREE: Master of Science
COLLEGE: Sciences and Humanities
DATE:MAY 2016
PAGES: 33
Bivariate geometric distribution is an extension to the univariate geometric
distribution that models the number of trials to get the first success. Thus
a bivariate distribution can be considered as a model for number of trials to
obtain two different but related events for the first time. Many statisticians
have studied different forms of bivariate geometric distribution. In this thesis, we
considered the form which is given by Phatak and Sreehari (1981). We estimated
the parameters using maximum likelihood estimation. We derived the deviances
as the goodness of fit statistics for testing the parameters corresponding to
reduced model, generalized linear model (GLM) and the deviance difference to
compare two models in order to determine which model fits the data well. To
determine the efficiency of our deviances, we simulated data using computer
software R. We found that our deviance for the reduced model with pair of
parameters of the bivariate geometric distribution worked well.