ABSTRACT
In 1956 a result due to Charles Stein was published that represented an important breakthrough in statistical estimation theory. Stein showed that the maximum likelihood estimator (MLE) for the mean of a multivariate normal distribution is inadmissible. This means that it is possible to construct an estimator with smaller risk than the MLE for the entire parameter space. James and Stein (1961) exhibited an estimator with risk uniformly smaller than that of the MLE. This estimator is now commonly referred to in the literature as the James-Stein estimator.
