Shrinkage estimation of the common location parameter of several exponentials

Estimation of the common location parameter of several exponentials is considered. Using samples from m independent exponential populations with common location parameter theta, and given a prior guess theta(0) of theta, several shrinkage estimators have been proposed that incorporate this prior information. We propose shrinkage factors by minimizing the mean squared error or utilizing the P-values obtained from combining certain independent statistics and tests. A simulation study is conducted to investigate the performance of the proposed estimators. It is found that the proposed estimators are effective in taking advantage of the available prior information.