非线性再生散度模型两种估计方法的比较  

A Comparison of Two Estimation Methods in Nonlinear Reproductive Dispersion Models

作　　者：左艳芳[1] 刘伟[2] 

机构地区：[1]昆明冶金高等专科学校社会科学与公共学院,云南昆明650033 [2]昆明冶金高等专科学校实践教学中心,云南昆明650033

出　　处：《云南民族大学学报（自然科学版）》2008年第4期320-323,350,共5页Journal of Yunnan Minzu University:Natural Sciences Edition

摘　　要：介绍以RDM为随机误差的广义线性模型-非线性再生散度模型的定义,分别考虑非线性再生散度模型的极大似然估计和Bayes估计,以模拟研究和实例分析对以上估计方法加以比较说明.Reproductive Dispersion Model （RDM） include a wide range of distributions, such as exponential family distribution, normal distribution, poissondistribution, etc, which are the extension and development of the results for generalized linear models, nonlinear regression models and exponential family nonlinear models. The definition of the NRDM is discussed, and both EM approach and Bayesian approach are developed for analyzing the model. In order to compare the efficiency of the two methods, a simulation study and a real example are presented to illustrate the methods.

关 键 词：非线性再生散度模型 Fisher得分算法 GIBBS抽样 

分 类 号：O212[理学—概率论与数理统计]