Summary
The concept of ignorance prior distribution is extended to the case of a hyperparameter. This leads to a procedure of formulating the partial ignorance of the original parameter. Its application to the estimation of the mean of a multivariate normal distribution with a particular hyperparameterized prior distribution of the mean leads to an improper prior distribution with the corresponding posterior mean very close to the James-Stein estimate.
Similar content being viewed by others
References
Akaike, H. (1978). A new look at the Bayes procedure,Biometrika,65, 53–59.
Brown, L. D. (1971). Admissible estimators, recurrent diffusions, and insoluble boundary value problem.Ann. Math. Statist.,42, 855–903.
Efron, B. (1978). Controversies in the foundation of statistics,Amer. Math. Monthly,85, 231–246.
Jeffreys, H. (1961).Theory of Probability, 3rd ed., Oxford University Press, London.
Stein, C. (1962). Confidence sets for the mean of a multivariate normal distribution,J. R. Statist, Soc., B,24, 265–296.
Stein, C. (1973). Estimation of the mean of a multivariate normal distribution,Proceeding of the Prague Symposium on Asymptotic Statistics, 345–381.
Strawderman, W. E. (1971). Proper Bayes minimax estimators of the multivariate normal mean.Ann. Math. Statist.,42, 385–388.
Takada, Y. (1979). Stein's positive part estimator and Bayes estimator,Ann. Inst. Statist. Math.,31, A, 177–183.
Additional information
The Institute of Statistical Mathematics
About this article
Cite this article
Akaike, H. Ignorance prior distribution of a hyperparameter and Stein's estimator. Ann Inst Stat Math 32, 171–178 (1980). https://doi.org/10.1007/BF02480322
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/BF02480322