Large-Sample Inference of EM Based on Maximum Likelihood Estimates

被引：0

作者：

Li, Yong ^{[1
]}

机构：

[1] Jilin Teathers Inst Engn & Technol, Coll Informat Engn, Changchun 130052, Peoples R China

来源：

MODERN TECHNOLOGIES IN MATERIALS, MECHANICS AND INTELLIGENT SYSTEMS | 2014年 / 1049卷

关键词：

EM algorithm; Loglikelihood; Information Matrix; Missing data analysis;

D O I：

10.4028/www.scientific.net/AMR.1049-1050.1343

中图分类号：

T [工业技术];

学科分类号：

08 ;

摘要：

EM algorithm is a very popular algorithm in missing data analysis. However, The variance of the estimator from EM is intractable. In this paper, we propose the supplemented EM algorithm for computing the variance that do not require computation and inversion of the information matrix.

引用

页码：1343 / 1346

页数：4