Empirical Bayes Estimation with Random Censoring

摘要

We study empirical Bayes estimation for exponential distributions based on randomly censored data. An empirical Bayes estimator ψ_n is constructed. The rate of asymptotic optimality of ψ_n is investigated. It is shown that under certain conditions, the regret of ψ_n converges to zero at a rate O(1n~4n/n~(1/2)), where n is the number of past data available when the present estimation problem is considered.