混合地理加权空间自回归模型的Profile GMM估计  

Profile GMM Estimation in Mixed Geographically Weighted Spatial Autoregressive Models

作　　者：马笑笑 王韶郡 魏传华[1] MA Xiaoxiao;WANG Shaojun;WEI Chuanhua(Department of Statistics,School of Science,Minzu University of China,Beijing 100081,China)

机构地区：[1]中央民族大学理学院统计学系,北京100081

出　　处：《应用数学》2023年第1期237-247,共11页Mathematica Applicata

基　　金：国家社会科学基金项目(21BTJ005)。

摘　　要：为了更好地同时考虑空间自相关性和空间异质性,本文研究一类空间自回归混合地理加权回归模型.基于Profile方法和广义矩(GMM)方法,构造了模型中未知空间自回归参数,常数回归系数和系数函数的两类Profile GMM估计.数值模拟结果表明所提出的估计在有限样本中表现良好.To deal simultaneously with spatial autocorrelation and spatial heterogeneity,this paper considers a mixed geographically weighted spatial autoregressive model.Based on the profile method and Generalized Method of Moments approach,two kinds of profile GMM estimators of the unknown spatial autoregressive parameter and constant regression coefficients,as well as nonparametric functions are proposed.Some simulations are conducted to examine the performance of our proposed methods and the results are satisfactory.

关 键 词：空间自回归 地理加权回归 广义矩估计 Profile方法 

分 类 号：O212.1[理学—概率论与数理统计]