Doctoral Theses

Survival Function Estimation Under Random Censoring.

2-28-1992

2-28-1993

Institute Name (Publisher)

Indian Statistical Institute

Doctoral Thesis

Degree Name

Doctor of Philosophy

Mathematics

Department

Theoretical Statistics and Mathematics Unit (TSMU-Kolkata)

Supervisor

Mukhopadhyay, Anis Chandra (TSMU-Kolkata; ISI)

Abstract (Summary of the Work)

Statistical methods for life data analysis are used to measure, compare and predict characteristics of the distribution of the time to some particular event of interest, often called failure after a length of time, called life time. Failure can occur at most once for an individual. Examples of failure time include the lifetimes of machine components in industrial reliability, the duration of strikes or periods of unemployment in economic studies, the time taken by subjects to complete specified tasks in psychological experiments, the lengths of tracks photographic plates in particle physics and the on survival time of patients in clinical trials. In life testing problem, reliability or medical follow up studies and other ficlds, the observations on the lifetime may not be possible for some sample units because of the occurence of other event (say loss). For instance in reliability studies this might happen because of a measure taken to avoid destructive life testing or because of limited availability of testing facilities. In medical follow up surveys, the patient often withdraws from the survey or the experimenter may not be able to make final contact before the death of a patient. In all such cases, say that the we observation is censor ed. In spite of this incompleteness of data owing to censoring, it is important to estimate the survival function of the lifetime of a unit from the data set observed or collected. Censoring may be random or fixed and in the present work we consider only the of random censoring which is case identified as one where the limits of observations set are values of another random variable, usually assumed to be distributed independently of the life time distribution of the unit. Again the unit may represent a system consisting of one or more components arranged in a certain way and the system life is dependent on component lives based on this arrangment. In our present work we have mainiy focused our attention to nonparametric estimation of survival functions of components under random censoring in a (a) two components parallel system and (b)k-1 out of k components system (k>2) with specia1 emphasis nn tun nut nf t.hree system.

Comments

ProQuest Collection ID: http://gateway.proquest.com/openurl?url_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:dissertation&res_dat=xri:pqm&rft_dat=xri:pqdiss:28842967

ISILib-TH192

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License.

DOI

http://dspace.isical.ac.in:8080/jspui/handle/10263/2146

COinS