Joe Padgett
Department of Statistics
University of South Carolina
Smooth Confidence Bands for the Survival Function from Randomly
Right-Censored Data
Randomly right-censored data arise often in industrial life testing and
medical trials. Asymptotic confidence bands for the survival function from
such data have been proposed by several authors. All of these bands are
based on the empirical estimator of the survival function, i.e. the
product-limit estimator. Here, asymptotic confidence bands are developed
from the kernel smoothed estimator of the survival function under the
general right-censorship model. These new bands are compared to empirical
bands by simulations, indicating that for small sample sizes the smooth
bands perform much better than the empirical counterparts.
Back to Colloquium Series