摘要
近年来空间自回归模型被广泛应用,然而它并没有考虑函数型自变量,在很多实际应用中,研究者可能会关注因变量的分位数情况,为此提出了函数型分位数空间自回归模型.首先利用函数型主成分分析方法估计非参数斜率函数,然后利用工具变量分位数回归方法估计参数.在一定的条件下,给出了参数和非参数估计的渐近性质.模拟研究表明工具变量分位数回归方法得到的估计可以减少估计的偏差,最后将模型应用到经济增长实际数据中得到所提出估计方法的良好表现.
Recently spatial autoregressive models have been widely used,but they do not consider functional independent variables.In many practical applications,researchers may pay attention to the quantile of the dependent variable,so a functional quantile spatial autoregressive model is proposed.Firstly,the nonparametric slope function is estimated by functional principal component analysis method,and then the instrumental variable quantile regression method is used to estimate the parameters.Under certain conditions,the asymptotic properties of parametric and non-parametric estimation are given.Simulation results show that the estimation obtained by the instrumental variable quantile regression method can reduce the estimation bias,and finally we apply the model to the actual data of economic growth to obtain the good performance of the proposed estimation method.
作者
刘高生
柏杨
LIU Gaosheng;BAI Yang(School of Sciences,Tianjin University of Commerce,Tianjin 300134;School of Statistics and Management,Shanghai University of Finance and Economics,Shanghai 200433)
出处
《系统科学与数学》
CSCD
北大核心
2023年第12期3361-3376,共16页
Journal of Systems Science and Mathematical Sciences
基金
天津市教委科研计划项目成果(函数型半参数空间自回归模型的统计推断及其应用)(2021SK140)资助课题.
关键词
分位数回归
函数型空间自回归模型
函数型主成分分析
工具变量
Quantile regression
functional spatial autoregressive model
functional principle component analysis
instrumental variable