Unit:
Τομέας Στατιστικής και Επιχειρησιακής ΈρευναςLibrary of the School of Science
Author:
Βαρελτζή Φωτεινή- Θεανώ
Supervisors info:
Φώτιος Σιάννης Λέκτορας, Νικόλαος Παπαδάτος Αναπλ. Καθηγ., Χαράλαμπος Δαμιανού Καθηγ.
Original Title:
Σύγκριση μοντέλων για την ανάλυση δεδομένων επιβίωσης με διακεκομμένες παρατηρήσεις πιθανόν πληροφοριακές
Summary:
In survival data analysis there is plenty of statistical models used to
interpret censored survival data. Our aim is to describe three different
statistical methods in order to face such data. The information that could be
extracted from censored data, as well as the number of censored observations
could be of great significance. In this thesis, we investigate the attitude and
results of these models on specific simulated data with different percentages
of right censored data each time, in order to make assumptions about the
suitability of each method. The inferences we make depend on criteria that have
already been declared.
Keywords:
Survival analysis, Informative censoring, Sensitivity analysis, Proportional hazards model, Inverse probability weighting
Number of pages:
VIII,119