Life models and parameter estimates in censored sample

LIFE MODELS AND PARAMETER ESTIMATES IN CENSORED SAMPLE Survival analysis which can be used in many fields such as engineering, medicine, insurance, economics, social sciences, is analyzed in this study on living things. Survival analysis can be defined as set of methods to measure survival time of an infected person under observation and determine effect of other factors on survival time. In this study various definitions of survival time are given and lifetime which is the most important element of this analysis is examined in detail. Definitions of basic concepts used in survival analysis such as life function, the possibility of sudden death (Hazard function), the distribution function and relations between them is explained. Censorship and censorship types is discussed in detail. In the fourth and fifth parts, Weibull distribution is examined for parameter estimates which are the main goal of this study and maximum likelihood estimators is focused for normal and censored samples. In the final part of the study data about ages and pregnancy period of 100 pregnant women is first examined for suitability of the Weibull distribution, then suitable censorship types is chosen for available data. In the event of censored sampling survival models and parameter estimates are prepared. Graphics of functions that gives info about life function, Hazard function that are important parts of survival analysis are prepared via Minitab 17 program. In this thesis it is aimed to determine parameter estimates that match with the life expectancy of the distribution. Keywords: Survival Analysis, Censoring, Censored Data, Parametric Life Models, Weibull Distribution, Parameter Estimates.