![]() | Only 14 pages are availabe for public view |
Abstract The exponential distribution is frequently used in lifetime data analysis, but its suitability is restricted to constant hazard (failure) rates. Lin et al. (1989) introduced the inverted exponen- tial distribution to overcome this problem. The generalized inverted exponential distribution was proposed as another useful two-parameter generalization of the inverted exponential dis- tribution (see Abouammoh and Alshingiti (2009)This lifetime distribution is capable of modeling diverse shapes of failure rates, and thus various shapes of aging criteria. In this thesis, the problem of estimation for generalized inverted exponential distribution have been studied under progressive Type-I censoring from Bayesian and non-Bayesian view- points. Maximum likelihood estimates and associated asymptotic confidence interval and boot- strap confidence interval have been derived for the unknown parameters of the generalized inverted exponential distribution. Based on Markov Chain Monte Carlo (MCMC), Bayes es- timates have been calculated using Metropolis-Hasting algorithm and the corresponding high- est posterior density credible interval estimates under non-informative and informative priors considering squared error loss function. Also, a discussion of how to select the values of hyper-parameters is taken into consideration based on past samples when informative prior is proposed. A real data set has been analyzed and a simulated study has been conducted to compare the performance of the various proposed estimators |