Abstract
dc:descriptionSurvival analysis is identical to meaning analysis of the period preceding the recording of an event. Asa term, it is primarily used in biomedical sciences, where the focus is on the observation of life ofpatients or laboratory animals. The analysis of the time leading up to an event record has also beenwidely used in the social sciences, where the focus is on events such as job change, marriage, birth ofchildren and so on. Additionally, the engineering sciences have contributed to the development ofsurvival analysis. In this scientific field is customary to use the term "reliability analysis", since the mainapplication is the modeling of the time required for the failure of machinery or electronics. The study ofall these scientific specialization is mostly consolidated in the area of survival analysis.It is easily understood that the description and modeling of life is the main issue in many aspects ofhuman life and activity. Therefore, the methodologies developed in survival analysis are a valuable toolfor researchers in various disciplines such as engineering, economics, medicine and other lifesciences.This thesis addresses issues related to continuous probability distributions that have direct applicationin survival analysis. In this context, the first chapter provides a brief historical overview of the researchfield of survival analysis, record the terminology and basic concepts of continuous distributions oflifetimes and describes the various characteristic functions and properties that they have. Additionally,ways to produce lifetime distributions, the analytical methodology for their study, and the mostcharacteristic distributions with fixed, unimodal or concave hazard functions, are presented in the samechapter.In the next four chapters, three different survival analysis models, as well as a family of lifetimedistributions and a four-parameter model produced from it, are introduced. For all models, variousproperties, such as ways of production, relations with other distributions, the monotony of the pdf, theirmoments, their entropies and the corresponding survival, hazard and mean residual life functions arestudied. The parameter estimation problem is treated by the maximum likelihood method, the observedand expected information tables are given, while the programming language for statistical applicationsR is used for calculating the estimators. Each of these chapters concludes with applications of fittingthe corresponding models to real datasets.The sixth chapter presents a review on beta generalized probability distributions. Their productionmethod is introduced and their basic mathematical properties are described. Additionally, variouslifetime distributions that belong to the beta generalized family that have been proposed so far in theliterature are presented, together with suggestions for future research in this field.By completing, a summary of the thesis in English language and the literature that was usedthroughout the research development are given.
Degree
thesis:*- Grantor dc:publisher
- University of Ioannina
- Year dc:date
- 2013
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Pappas, Vasileios
- Παππάς, Βασίλειος
Subjects
dc:subject × 20- Μαθηματική στατιστική
- Ανάλυση επιβίωσης
- Κατανομές χρόνου ζωής
- Αξιοπιστία
- Έλεγχος χρόνου ζωής
- Συνάρτηση κινδύνου
- Στατιστική συμπερασματολογία
- Προσαρμογή πραγματικών δεδομένων
- Mathematical statistics
- Survival analysis
- Lifetime distributions
- Reliability
- Life testing
- Hazard function
- Statistical inference
- Real data fit
- Φυσικές Επιστήμες
- Μαθηματικά
- Natural Sciences
- Mathematics
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/29368