一类混合正态分布参数估计的EM算法和数据扩张

EM algorithm and data augmentation for a class of Gaussian mixtures

从解析的角度得到混合正态分布参数的ML估计是很困难的.通过引入"缺失数据",利用Little和Schluchter(1985)提出的处理分类和连续数据的方法,给出具有相同协方差阵的混合正态分布参数估计的EM算法和数据扩张算法.

It is difficult to use analystic methods to determine ML estimation for Gaussian mixtures.In this article,after introducing missing data,EM algorithm and data augmentation are given for parameters estimations of Gaussian mixtures with the common covariance matrix.