|
Document Type |
: |
Thesis |
|
Document Title |
: |
A New Generalization of T-X Family of Distributions تعميم جديد لعائلة التوزيعات تي – اكس |
|
Subject |
: |
Faculty of Science |
|
Document Language |
: |
Arabic |
|
Abstract |
: |
Statistical distributions play an important role around many world phenomena. They can be applied in many fields, such as computer sciences, engineering, demography and medical sciences. However, some of the existing classical distributions cannot describe and fit the complex behavior of real data. Consequently, more adaptable distributions are continuously required in order to improve data fitting. Regression analysis is also considered an important tool for analyzing relationships between variables. More specifically, regression analysis deals with the study of the dependence of one variable on single or multiple independent variables.
Thus, the main purpose of this thesis is to introduce a new method for generating a family of distributions that are more flexible at fitting different data. Furthermore, these distributions are generalized to examine regression models. Specifically, this thesis proposed three novel distributions namely, the new generalized exponentiated generalized inverted Kumaraswamy Gompertz distribution, the new generalized exponentiated Frechet Weibull distribution and the new generalized exponentiated type II Topp Leone Weibull Burr XII distribution. Further, two new location-scale regression models are constructed, one based on the new generalized exponentiated Frechet Weibull and the other based on the new generalized exponentiated type II Topp Leone Weibull Burr XII distributions.
For each of these newly developed distributions, some of the statistical and mathematical properties are discussed. In addition, their parameters are estimated using the method of maximum likelihood. Some Monte Carlo simulations and several real data sets are examined in order to illustrate the flexibility of the introduced distributions. Furthermore, for each regression model, the maximum likelihood method is applied to estimate the models parameters. A number of simulation studies have been conducted and real data sets have been examined to evaluate the usefulness of the suggested models.
In conclusion, the overall results indicated that the proposed models offer an increased degree of flexibility that is capable of analyzing real data in a variety of applications. |
|
Supervisor |
: |
Dr. Hadeel Klakattawi |
|
Thesis Type |
: |
Master Thesis |
|
Publishing Year |
: |
1444 AH
2023 AD |
|
Co-Supervisor |
: |
Prof. Dr. Lamya Baharith |
|
Added Date |
: |
Monday, July 17, 2023 |
|
Researchers
| عائشه عبده خرمي | Khormi, Aisha Abdou | Researcher | Master | |
|