Inferences on the mixture model with type-II hybrid censored data

Abstract

In many practical situations, the population under study consists two or more sub populations. More so, the presence of multiple modes of failuremay create a heterogeneous component population. The present study deals with the statistical inferences on the mixture of two exponential distributions with Type-II hybrid censored data. The maximum likelihood estimators and confidence intervals of the parameters are developed. Bayes estimates and the corresponding highest posterior density credible intervals of the unknown parameters are obtained using Jeffrey’s invariant priors on the unknown parameters, and by using Gibbs sampler. The performances of the methods of estimation are compared with Monte Carlo simulations.