Analysis of Type-Ⅱ censored data from the Kies distribution with classical and generalized approaches

This paper investigated statistical inference for the Kies distribution under the Type-Ⅱ censoring scheme; specifically, classical and alternative generalized inferential approaches were proposed for parameter estimation. From the classical likelihood perspective, maximum likelihood estimators for unknown parameters were derived, and the existence and uniqueness of these estimators was also established. The corresponding asymptotic confidence intervals were constructed based on the observed Fisher information matrix. For comparison, an alternative generalized estimation method was conducted based on the constructed pivotal quantities. Finally, the performance of different estimation methods was evaluated via extensive simulation studies, and meanwhile, two real-world data examples were presented to illustrate the applications of the proposed methods.