|国家科技期刊平台
首页|期刊导航|统计与决策|半参数空间自回归变系数模型的统计推断

半参数空间自回归变系数模型的统计推断OA北大核心CHSSCDCSTPCD

Statistical Inference of Semi-parametric Spatial Autoregressive Variable Coefficient Model

中文摘要英文摘要

半参数空间自回归变系数模型因同时考虑了因变量的空间自相关性和回归关系的空间异质性而具有广阔的应用前景.文章针对此模型,基于轮廓拟极大似然估计,提出了常值系数和局部系数的显著性检验方法,以辨识模型中的零值系数,从而更深入地理解回归关系的变化特征;同时,也给出了局部检验可能涉及的多重检验的解决方法.模拟实验验证了所提检验方法的有效性,而基于美国波士顿房屋价格数据的实证分析验证了检验方法的应用效果.

The semi-parametric spatial autoregressive variable coefficient model has a broad application prospect because it takes into account both the spatial autocorrelation of dependent variables and the spatial heterogeneity of regression relations.Based on contour quasi-maximum likelihood estimation,this paper proposes a significance test method of constant coefficient and local coefficient to identify the zero-value coefficient in the model,so as to understand the change characteristics of regression re-lations more deeply.At the same time,is also given the solution of multiple tests which may be involved in local tests.Simulation experiments verify the validity of the proposed test method,and the empirical analysis based on the house price data of Boston in USA verifies the application effect of the test method.

陈凤;刘嘉慧

重庆交通大学 数学与统计学院,重庆 400074西安交通大学 管理学院,西安 710049

经济学

空间非平稳性空间自相关性统计推断多重检验

spatial non-stationarityspatial autocorrelationstatistical inferencemultiple test

《统计与决策》 2024 (004)

17-22 / 6

国家社会科学基金重点项目(20AGL004)

10.13546/j.cnki.tjyjc.2024.04.003

评论