Exact Likelihood Inference for an Exponential Parameter under a Multi-Sample Type-II Progressive Hybrid Censoring Model

Abstract

This paper introduces multi-sample Type-II progressive hybrid censoring which allows to incorporate data from $k\ge 2$ independently conducted Type-II (progressive) hybrid censored experiments for the first time. Assuming exponentially distributed lifetimes with mean $\vartheta ,$ the maximum likelihood estimator of $\vartheta$ is obtained and its exact distribution is established. Moreover, we propose an exact confidence interval as well as two asymptotic confidence intervals for the unknown mean. The results are illustrated by simulations and a data set.

KEYWORDS AND PHRASES:

• Confidence intervals
• exponential distribution
• multi-sample hybrid censoring
• multi-sample Type-II progressive hybrid censoring
• progressive hybrid censoring

Acknowledgment

The authors are grateful to two reviewers for constructive comments which led to an improved version of the manuscript.