Characterization and identification of probability distributions.
Date
1991
Authors
Rahman, Md. Shafiqur.
Journal Title
Journal ISSN
Volume Title
Publisher
Dalhousie University
Abstract
Description
Let P(x; $\theta$) be the probability mass function or probability density function of a random variable X where $\theta\varepsilon\ \Re\sp{\rm p}$, p being finite. Using the first k (k $\ge$ p) raw or central moments of this distribution we eliminate the p parameters in $\theta$ and obtain a moment relation in k moments. We derive the raw and the central moment relations for a number of discrete and continuous distributions. These moment relations are used as criteria to characterize a distribution. In general the present method is effective. But there are some special situations, where the moment relations of two or more distributions are same or one particular moment function takes same value for two or more distributions. In such a situation we propose two moment ratios as extra criteria for deciding among them. These ratios are also useful in approximating the Neyman type A and the Generalized Poisson distribution by the Negative Binomial distribution. We can identify a distribution by using the ratios of the co-efficients of the recurrence relations obtained from its generating function.
Subsequently, a special class of the Exponential family of distributions named the family of Transformed Chi-square distributions is defined. Explicit expressions for the MVUE with MV of a function of the parameter of this family are given. The critical region and the power function for various tests of hypotheses for the parameter of this family are also obtained. An identification procedure with probability of correct identification is discussed in detail.
Thesis (Ph.D.)--Dalhousie University (Canada), 1991.
Subsequently, a special class of the Exponential family of distributions named the family of Transformed Chi-square distributions is defined. Explicit expressions for the MVUE with MV of a function of the parameter of this family are given. The critical region and the power function for various tests of hypotheses for the parameter of this family are also obtained. An identification procedure with probability of correct identification is discussed in detail.
Thesis (Ph.D.)--Dalhousie University (Canada), 1991.
Keywords
Statistics.