Estimation of Inverse Lindley Distribution Using Product of Spacings Function for Hybrid Censored Data

This article presents different estimation procedure for inverse Lindley distribution for Type-I hybrid censored data. We have obtained the parameter estimate under both the classical and Bayesian paradigm. In the classical set up, method of Maximum Likelihood(ML) and Maximum Product of spacings (MPS) estimates are obtained along with 95% asymptotic confidence interval. Bayesian estimation is implemented under the assumption of squared error loss function. An alternative Bayesian procedure is also proposed by incorporating the sample information through the spacings function instead of likelihood function. The Bayes estimates are computed using Markov Chain Monte Carlo (MCMC) technique due to their implicit nature. Highest posterior density (HPD) intervals based on these MCMC samples are evaluated and compared in terms of simulated risks. Further, a real data of 72 guinea pigs, infected with tuberculosis is analysed to justify the suitability of the afore-said estimation techniques under the specified censoring scheme.