Robust estimators of the 2-parameter gamma distribution

The available estimators for parameters of the gamma distribution are moment estimators, maximum-likelihood estimators, and approximations to the maximum-likelihood estimators. These estimators are not suitable for small samples; however, they are still being used at the present time. The proposed robust estimators for scale and shape parameters are more suitable for small samples. They have RMS (root-mean-square) errors that are considerably smaller than those of the other estimators. In addition, they are easier to calculate, and are therefore appropriate in many applications.< >