参数的变换矩估计方法

A MOMENT ESTIMATOR BASED ON THE PROBABILITY INTEGRAL TRANSFORMATION

概率积分变换是统计分析中一个重要方法，它在拟合优度检验中得到了良好的应用．本文将这一思想应用于参数估计问题，提出一般参数估计的变换矩方法研究表明，这种方法既有较高的效率，又有良好的稳健性．文中证明了变换矩估计的存在性，相合性和渐近正态性，同时给出了影响函数和崩溃点，并具体分析了几个常见分布族的结果．

This paper presents a new method for estimating unknown parameters in populations. The new method is based on the probability integral transformation and matching the moments of the transformed sample to the moments of uniform distribution on (0, 1).The properties, including consistency, asymptotical normality and robustness, of the proposed estimator are discussd. Some examples are presented to illustrate the method.