Abstract
The paper revisits the entropy-based efficiency of the type-I censored sample, which was addressed by several previous works. The main purpose of this work is to provide a comprehensive definition of the efficiency function and give a general proof that the entropy of a censored sample is always less than that of the complete sample for any probability distribution and at any point of censoring. A simulation study is performed to validate our results, and a real-data example is reevaluated.