固定效应部分线性单指标面板模型的惩罚二次推断估计  

Penalized Quadratic Inference Estimation for Partially Linear Single-Index Panel Model with Fixed Effect

作　　者：陈建宝 丁飞鹏 CHEN JIANBAO;DING FEIPENG(College of Mathematics and Informatics,Fuvjian Norrmal University,Fuzhou 350117,China;School of Economics,Zhejiang University of Finance and Economics,Hangzhou 310018,China;School of Mathematics and Statistics,Jiangxi Normal University,Nanchang 330022,China)

机构地区：[1]福建师范大学数学与信息学院,福州350117 [2]浙江财经大学经济学院,杭州310018 [3]江西师范大学数学与统计学院,南昌330022

出　　处：《应用数学学报》2021年第1期121-144,共24页Acta Mathematicae Applicatae Sinica

基　　金：国家社会科学基金项目(16BTJ018);教育部人文社会科学重点研究基地重大项目(15JJD790029);教育部人文社会科学研究项目(13YJA9100002);福建省自然科学基金项目(2017J01396,2018J05002);福建省高校科技创新团队和福建师范大学创新团队项目(IRTL1704);国家自然科学基金(71961011);江西高校人文社科基金(TJ19203)资助。

摘　　要：单指标面板模型已广泛应用于各学科领域的研究中,其估计方法较为丰富,然而鲜有估计方法将个体内的相关性考虑在内.基于此,本文研究了一类个体内存在相关性的固定效应部分线性单指标面板模型,采用惩罚二次推断函数法和LSDV法相结合的方法对模型进行估计,证明了所得估计量的一致性和渐近正态性.Monte Carlo模拟结果显示其具有优良的有限样本表现,并将该估计技术应用于实际数据分析中.Single-index panel models have been widely used in a variety of research fields.They have many different estimation methods.However,few methods consider the case of correlation in subjects.Therefore,this paper studies a partially linear single-index panel model with fixed effect which has correlation in subjects.We estimate the model using combination of penalized quadratic inference function and LSDV.The consistency and asymptotic normality of the estimators are derived.Meantime,Monte Carlo simulation shows that the finite sample performances of the proposed estimation method are still very good.Furthermore,the proposed estimation technique is illustrated in the analysis of a real data set.

关 键 词：固定效应 部分线性单指数面板模型 惩罚二次推断估计 LSDV法 

分 类 号：O212.1[理学—概率论与数理统计]