Abstract
This paper studies a monotone empirical Bayes test δ * n for testing H 0:θ≤θ 0 against H 1:θ>θ 0 for the positive exponential family f(x|θ)=c(θ)u(x)exp (−x/θ), x>0, using a weighted quadratic error loss. We investigate the convergence rate of the empirical Bayes test δ * n . It is shown that the regret of δ * n converges to zero at a rate O(n −1), where n is the number of past data available when the present testing problem is considered. Errors regarding the rate of convergence claimed in Gupta and Li (J. Stat. Plan. Inference, 129: 3–18, 2005) are addressed.
Similar content being viewed by others
References
Gupta, S.S., Li, J.: On empirical Bayes procedures for selecting good populations in a positive exponential family. J. Stat. Plan. Inference 129, 3–18 (2005)
Gupta, S.S., Liang, T.: Selecting good exponential populations compared with a control: a nonparametric empirical Bayes approach. Sankhya, Ser. B 61, 289–304 (1999)
Gupta, S.S., Liese, F.: Asymptotic distribution of the conditional regret risk for selecting good exponential populations. Kybernetika 36, 571–588 (2000)
Liang, T.: On Singh’s conjecture on rate of convergence of empirical Bayes tests. J. Nonparametr. Stat. 12, 1–16 (1999)
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Liang, T. On a monotone empirical Bayes test in a positive exponential family. J. Appl. Math. Comput. 32, 97–109 (2010). https://doi.org/10.1007/s12190-009-0235-8
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s12190-009-0235-8
Keywords
- Asymptotically optimal
- Empirical Bayes
- Positive exponential family
- Rate of convergence
- Weighted error loss