J. de Uña Álvarez
Doubly truncated data appear in Epidemiology, Survival Analysis, Reliability Engineering, Astronomy and Economics, among other fields. With doubly truncated data inference is typically performed through the Efron-Petrosian nonparametric maximum-liklihood estimator, a non-explicit estimator with complicated asymptotic properties. In this talk I will introduce some recent developments in goodness-of-fit testing under double truncation. This includes testing for a fully or parametrically specified distribution of the target variable; testing for a parametric model for the truncation couple (which is needed in the semiparametric double truncation model); testing for an ignorable sampling bias on the target outcome; and testing for regression models. Theoretical results will be discussed. Simulation studies and real data illustrations will be given.
Keywords: bootstrap approximation, Interval sampling, random truncation, Survival Analysis, nonparametric statistics
Scheduled
Invited Session Recent advances in goodness-of-fit and k-sample tests
June 10, 2022 10:10 AM
A13