文章摘要
范巧,郭爱君.一种新的基于全息映射的面板时空地理加权回归模型方法[J].数量经济技术经济研究,2021,(4):120-138
一种新的基于全息映射的面板时空地理加权回归模型方法
A New Geographically and Temporally Weighted Regression Model for Panel Data Based on Holographic Mapping
  
DOI：
中文关键词: 面板时空地理加权回归模型  时空权重矩阵  同伴效应  全息映射
英文关键词: Geographically and Temporally Weighted Regression Model for Panel Data （PGTWR）  Spatiotemporal Weight Matrix  Peer Effects  Holographic Mapping
基金项目:本文获得2019年度重庆市社会科学规划项目（2019YBJJ045）和2019年度重庆市留学归国人员创新创业支持计划人选项目（cx2019112）的资助。
作者单位
范巧 兰州大学经济学院 
郭爱君 兰州大学经济学院 
中文摘要:
      研究目标：架构适应面板数据分析、基于全息映射的时空地理加权回归模型分析范式。研究方法：基于近邻局部点对目标分析局部点的全息映射构建了适应面板数据空间计量局部分析的内生时空权重矩阵；在此基础上基于局部点参数估计与模型整体性质分析的分解，系统地架构了面板时空地理加权回归模型方法；还基于索洛余值法与面板时空地理加权回归模型方法的结合，对1990~2018年中国省级层面的全要素生产率及其增长率进行了重估。研究发现：基于全息映射的面板时空地理加权回归模型方法全面地解析了空间局部点之间影响效应的直接路径和间接路径；既考察了空间近邻局部点的同伴效应，也考虑了空间局部点自身的内生动力；同时，基于最优空间带宽和最优时间带宽纳入有效的近邻局部点，使得局部点空间依赖的规律性和异质性分析更为精准。研究创新：建立了新的基于全息映射的面板时空地理加权回归模型方法，编写了一整套基于MATLAB R2020a的标准化代码。研究价值：推动了面板时空地理加权回归模型方法的理论研究，并为面板数据空间计量局部分析的应用研究提供了标准化代码。
英文摘要:
      Research Objectives：Construct a new geographically and temporally weighted regression model for panel data （PGTWR） based on the holographic mapping. Research Methods：This paper constructed an endogenous spatiotemporal weight matrix based on the holographic mapping from the adjacent points to the target points，and also constructed the PGTWR model by two steps，including parameters estimation of local points and the overall model property analysis. On this basis，this paper reassessed the total factor productivity and its growth rate at the provincial level in mainland China during 1990~2018 using PGTWR and the Solow Residual Method. Research Findings：The PGTWR model based on holographic mapping can more accurately analyze the spatial and temporal relationships among spatial local points. In the PGTWR analysis，each local point and its effective adjacent points are chosen by the optimal spatial and temporal bandwidth to study the laws of spatiotemporal dependence and its heterogeneity. Also，The PGTWR model not only considers the peer effect from the adjacent points but also the endogenous power from each local point itself. Research Innovations：Promotes a new PGTWR analysis based on holographic mapping，and provides a set of corresponding standardized code based on Matlab R2020a. Research Value：Makes important marginal contribution for theoretical researches of the PGTWR analysis，and provides standardized Matlab code for the application research of the spatial econometrical local analysis with panel data.
查看全文