一类数据不完整情况下的概率分布参数估计方法

A Class of Methods to Estimate Parameters from Incomplete Data

提出了一类数据不完整情况下的参数估计方法 ,这类方法以分布函数或其各阶导数和频率间的关系为基础 ,并利用最小二乘法建立求解未知分布参数的方程组。另外 ,针对一些常见概率密度函数的特点 ,文中通过合理的近似 ,提出了一种简化的参数求解方法。实例表明本文方法求解出的未知分布参数精度较高 ,对于一些常见的概率分布函数 ,简化方法的求解过程相当简单。

In this paper a class of methods are presented to estimate the parameters from incomplete data.These methods are based on the relation between the distribution functions or its derivertives and the frequency.The least square thorem is used to derive general equations to solve unknown parameters.In addition,a simplfied method is discussed for the relation between the density function and frequency.Examples prove that the parameters estimated by these methods are more accurate,and simplified method can easliy estimate the parameters on the occasion of common distribution functions.