Aristotle University Of Thessaloniki (AUTH)
Συμβολή στη δειγματοληπτική ανάδειξη μοντέλων κατανομών με χρήση του συντελεστή μεταβλητότητας
Abstract
dc:descriptionThe aim of the present thesis is the examination of the behaviour of the Coefficient of Variation and its contribution to emerging distribution models via Sampling. The first two chapters describe what is known so far and review previous work done in the domain of the thesis. More precisely, the introductory chapter gives some definitions of basic concepts, which are needed as a basis for the subsequent chapters. The second chapter describes the concept of the Coefficient of Variation and its properties. Confidence intervals are presented based on the type of distribution and appropriate statistical hypothesis tests are given. The concepts of the Weighted Coefficient of Variation and the Multivariate Coefficient of Variation are also described, as well as unbiased estimators of the Coefficient of Variation. At the end of the chapter, we present some uses of the Coefficient of Variation and the Weighted Coefficient of Variation in various scientific fields. The following chapters are the originality of the thesis. A new function of polynomial form is introduced in the distribution theory, with a real number ν as a degree, which is calculated solely from the sample Coefficient of Variation. That new function, based on data distribution, has a specific formula and type, which can be either symmetric, ascending or descending type and approximates known or unknown distributions or emerges new ones. Moreover, the parallel theoretical and sampling study leads to the emergence of the correct or more nearly correct model of a distribution. Given that the degree obeys specific restrictions, which lead to a restricted Coefficient of Variation, information for the coefficient of variation is obtained of the distributions approximated by the new function. The values of the Coefficient of Variation are of great interest to researchers, who will identify the type of distribution via sampling and therefore will know the expected value of the Coefficient of Variation. If the value of the sample Coefficient of Variation moves away from the expected value, then the researcher should give an answer to this and draw conclusions depending on his scientific field. In the chapters 3, 4 and 5, the polynomial approach of the probability density functions (pdf) of symmetric type, ascending form and descending type, respectively, of a continuous random variable X, is attempted. The sixth chapter is devoted entirely to the contribution of the Coefficient of Variation to emerging distribution models via Sampling. Special forms of functions are described and representative examples are displayed. In the seventh chapter, a brief reference is made to the case of the discrete random variable X and we provide the answer to the question whether or not the sample Coefficient of Variation is a good estimator of the population Coefficient of Variation
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2018
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Papatsouma, Ioanna
- Παπατσούμα, Ιωάννα
Subjects
dc:subject × 12Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/46247