|
Abstract:
|
In this talk, I will review some results about large sample properties of the posterior distribution of the survival function with censored data when neutral to right processes priors are used. In particular, I will focus on the Bernstein von-Mises theorem, which implies that the posterior distribution is asymptotically equivalent to the sampling distribution of the MLE. First, I will explain sufficient conditions for the BvM theorem for right censored data. Then, I will talk about the BvM theorem for doubly censored data. Finally, I will discuss some open questions including interval censored data and frailty model.
|
