Inadmissibility of Common Estimators of Normal Covariance Matrix

In this paper, we give a new estimator improving the best invariant estimator of normal variance, and prove both common estimators of normal covariance and covariance matrix are inadmissible, when the loss functions are quadratic.