Abstract

We propose goodness-of-fit tests for Birnbaum-Saunders distributions with type-II right censored data. Classical goodness-of-fit tests based on the empirical distribution, such as Anderson-Darling, Cramer-von Misses, and Kolmogorov-Smirnov, are adapted to censored data, and evaluated by means of a simulation study. The obtained results are applied to real-world censored reliability data.