Reliability estimation for the inverted exponentiated Pareto distribution
Rani Kumari, Yogesh Mani Tripathi, Rajesh Kumar Sinha, Liang Wang
Abstract
We consider estimation of reliability in a multicomponent system when data are observed under Type-II censoring. Various estimates of this parametric function are derived when stress and strength (SS) variables follow inverse exponentiated distributions with a common scale parameter. We first obtain maximum likelihood estimate of the reliability. Then, approximate confidence intervals are obtained based on asymptotic theory. Further useful pivotal quantities are constructed and in sequel alternative estimates of the reliability are derived. The case where all parameters of SS components are unknown is also studied and various estimates for the reliability are proposed. Equivalence testing between model parameters is discussed as well. Performance of all estimates is compared using simulations and comments are derived. We analyze two real data sets for illustration purposes.