ABSTRACT
Survival data analysis tackles the problem of modeling observations of time to event. A more comprehensive treatment of Bayesian survival analysis can be found in J. G. Ibrahim et al. Survival data are often censored. For example, patients may experience the event during the study period, in which case the time to event is observed, or after the study has been finished, in which case the event time is unobserved. The Kaplan-Meier estimator provides a non-parametric estimate of the survival curve. Under some circumstances, survival may be similar for certain patients. For example, members of a family may be resistant to certain types of cancer or pieces manufactured at a particular factory may have a lower quality that decreases survival time. Frailty models are essentially mixed-effects models for survival data. Sometimes survival data are clustered or grouped and observations are not independent.
