An Empirical Bayes Improvement of Common Shrinkage Estimators

• Activity Number: 583
• Type: Invited
• Date/Time: Wednesday, August 3, 2016 : 2:00 PM to 3:50 PM
• Sponsor: International Indian Statistical Association
• Author(s): Eitan Greenshtein* and Ya'acov Ritov and Ariel Mansura
• Companies: Central Bureau of Statistics, Israel and Hebrew University/University of Michigan and Bank of Israel
• Keywords: empirical Bayes ; non-parametric

Abstract

We introduce a general method for improving the performance of common estimators of a Gausian mean vector, such as, Stein's estimator, Kalman filter, and many more. The improvement is asymptotical as the dimension grows, it is based on non-parametric empirical Bayes ideas and techniques.