半变系数固定效应面板数据模型研究PM_(2.5)与气候因素之间的动态关系  

Semi-Varying Coefficient Fixed-Effect Panel Data Model Explores the Dynamic Relations between PM_(2.5) and the Meteorological Factors

作　　者：胡雪梅 卢婵婵 杨艳林 HU Xuemei;LU Chanchan;YANG Yanlin(School of Mathematics and Statistics,Chongqing Key Laboratory of Social Economy and Applied Statistics,Chongqing Technology and Business University,Chongqing,400067,China;Research Center for Economy of Upper Reaches of the Yangtse River,Chongqing Technology and Business University,Chongqing,400067,China)

机构地区：[1]重庆工商大学数学与统计学院,经济社会应用统计重庆市重点实验室,重庆400067 [2]重庆工商大学长江上游经济研究中心,重庆400067

出　　处：《应用概率统计》2023年第1期53-72,共20页Chinese Journal of Applied Probability and Statistics

基　　金：supported by the Science and Technology Research Program of Chongqing Education Commission(Grant No.KJZD-M202100801);the Fifth Batch of Excellent Talent Support Program of Chongqing Colleges and University(Grant No.68021900601);the Natural Science Foundation of CQ CSTC(Grant No.cstc.2018jcyjA2073);the Program for the Chongqing Statistics Postgraduate Supervisor Team(Grant No.yds183002);the Chongqing Social Science Plan Project(Grant No.2019WT59);the Open Project from Chongqing Key Laboratory of Social Economy and Applied Statistics(Grant No.KFJJ2018066);the Mathematic and Statistics Team from Chongqing Technology and Business University(Grant No.ZDPTTD201906).

摘　　要：本文提出了一个半变系数固定效应面板数据模型来探索五个城市—–北京、成都、广州、上海和沈阳的PM_(2.5)与五个气候因素:累积风速、气压、露点、温度和每小时降雨量之间的动态关系,其中,PM_(2.5)是决定能见度的关键因素,对公众健康构成严重威胁.随后,我们将多元局部线性拟合、转换方法与剖面似然相结合,建立了参数向量和变系数函数向量的半参数固定效应估计.最后,展示了这五个城市2015年的估计动态关系.所提出的方法也可推广到经济和金融等其他领域的面板数据分析.In this paper we propose a semi-varying coefficient fixed effect panel data model to explore the dynamic relations between PM_(2.5) and the five meteorological factors:the cumulative wind speed,air pressure,dew point,temperature and hourly precipitation for the five cities:Beijing,Chengdu,Guangzhou,Shanghai and Shenyang,where PM_(2.5) is a key factor to determine visibility and a serious threat to public health.Then,we combine multivariate local linear fitting,transformation technique with profile likelihood,and establish semi-parametric fixed effects estimators for both parameter vector and varying-coefficient function vector.Finally,we exhibit their estimated dynamic relationships for the five cities in 2015.The proposed procedure can also be generalized to panel data analysis in other fields such as economy and finance.

关 键 词：气候因素 PM_(2.5)污染 半变系数固定效应面板数据模型 半参数固定效应估计 

分 类 号：O212.1[理学—概率论与数理统计]