This article discusses the parameter estimation problem in the two-parameter exponential distribution based on complete sample, type I censoring sample and random censorship sample, respectively. Employing linear regression method, Bayes approach and maximum likelihood method, we propose several kinds of estimators for the scale parameter and the location parameter and make comparisons between those estimators under each situation and also compare them with respective uniformly minimum variance unbiased estimator (UMVUE). The simulation results show that under the complete case the UMVUE performs the best, the Bayes estimator is the best in the type I case and the Similar uniformly minimum variance unbiased estimator(SUMVUE) is the best in the random censorship case.
|