Bayesian and classical analysis of time-to-failure model with comparison of uninformative priors

Rossman,et al. (1998) describe the relationship between bayesian and classical estimation using the continuous uniform distribution. in this paper,we provide the comparison of uninformative (jeffreys and uniform) priors for the parameter of the exponential model for time-to-failure data. the bayesian and classical analysis of the model is also presented here. our comparison is based upon the posterior variance,the bayesian point and interval estimates,the coefficients of skewness of the posterior distribution and the posterior predictive distribution. the bayes estimators under the different loss functions are also compared. © 2010 pakistan journal of statistics.