Eksik gözlemli tekrarlı ölçüm tasarımlarında veri analiz yöntemleri

---

Thesis Type: Postgraduate

Institution Of The Thesis: Gazi University, Fen Bilimleri Enstitüsü, İSTATİSTİK ANABİLİM DALI, Turkey

Approval Date: 1997

Thesis Language: Turkish

Student: MELTEM EKİZ

Supervisor: HAMZA GAMGAM

Abstract:

(M. Sc. Thesis) Meltem EKİZ GAZI UNIVERSITY INSTITUTE OF SCIENCE AND TECHNOLOGY November 1997 ABSTRACT In the repeated measurements and time-dependent data, it's frequently the case that some of the individuals couldn't have been measured on the concerned variable at one or more of the different time points. In this study; the data obtained from this kind of sample units, that are called "Missing observations", were added to the fully-classified data in order to estimate the cell probabilities. Therefore, the Multiple and Single Sample Models that makes the related hypotheses testing possible are investigated and compared both on partially and fully classified data. At last, the differences on the related estimators have been investigated for the application data, on with or without missing observations trough the method Weighted Least Squares. Science Code : 406.02.01 Key Words : Repeated Measurements, Maximum Likelihood Function, Contingency Tables, Missing Observations, The Weighted Least Squares Method, Fully and Partially Classified Data Page Number : 73 Adviser : Assoc. Prof. Hamza GAMGAM