半参数回归模型拟极大似然估计的弱相合性  

Weak Consistency of Quasi-maximum Likelihood Estimators in a Semiparametric Regression Model

作　　者：胡宏昌[1] 徐侃[1] 陈琴[1] 

机构地区：[1]湖北师范学院数学系,湖北黄石435002

出　　处：《工程数学学报》2008年第6期1081-1086,共6页Chinese Journal of Engineering Mathematics

基　　金：湖北省高校优秀中青年科技创新团队资助项目;湖北省教育厅重点项目(D200522002)

摘　　要：本文考虑一类固定设计的半参数回归模型,其误差为一阶自回归时间序列。用权函数及拟极大似然估计方法得到了一些参数及非参数的拟极大似然估计量,在适当的条件下,研究了它们的弱相合性,从而丰富了该类半参数回归模型的估计理论与方法。This paper considers a semiparametric regression model with deterministic design points, whose errors are the first order autoregressive time series. Some parametric and non-parametric quasimaximum likelihood estimators are obtained by using the weight function and quasi-maximum likelihood methods. Under proper conditions, consistencies of these estimators in probability are established, which enrich existing estimation theories and methods for semiparametric regression models.

关 键 词：半参数回归模型 AR(1)时间序列 拟极大似然估计 弱相合性 

分 类 号：O212.1[理学—概率论与数理统计]