A New Closed Form Estimator for Beta Distribution
Institute of Fundamental Sciences, Massey University, Palmerston North
In this study, we propose a new closed form estimator for two-parameter beta distribution. It is derived from the score equation of generalized beta distribution. The asymptotical properties of this estimator is investigated. It surprisingly resembles MLE in terms of mean square error and we further explore the linkage between the new estimator and MLE. With its closed form, it is very easy to develop some improved estimators such as one step estimator and jackknife estimator.