摘要
文章在响应变量随机缺失下研究非线性均值方差模型的参数估计问题。基于回归插补和随机回归插补两种缺失插补方法以及结合Gauss-Newton迭代计算算法给出该模型中未知参数的极大似然估计。并通过对两个随机模拟例子实际例子的研究分析,结果都表明了所提出的模型与统计方法具有可行性和实用性。
This paper studies the parameter estimation of nonlinear mean and variance models in the case of randomly missing response variable. Based on two imputation methods of regression imputation and stochastic regression imputation, and com bining with the Gauss-Newton iterative algorithm, the paper gives the maximum likelihood estimation of the unknown parameters in the model. Meanwhile, the paper conducts research analysis on two random simulation examples and practical data, and gets the result that the proposed model and statistical method are feasible and practical.
出处
《统计与决策》
CSSCI
北大核心
2017年第19期10-14,共5页
Statistics & Decision
基金
国家自然科学基金资助项目(11301485
11526188
11171105)
2016年度浙江省统计研究课题
