响应变量缺失下部分线性EV模型的异方差检验  

Heteroscedasticity Test for Partial Linear EV Model with Missing Response Variables

作　　者：刘锋 贺璟 高伟强 付馨苇 康新梅 LIU Feng;HE Jing;GAO Weiqiang;FU Xinwei;KANG Xinmei(College of Science,Chongqing University of Technology,Chongqing,400054,China)

机构地区：[1]重庆理工大学理学院,重庆400054

出　　处：《应用概率统计》2021年第4期346-360,共15页Chinese Journal of Applied Probability and Statistics

基　　金：国家自然科学基金项目(批准号:11471060、11871124)资助.

摘　　要：本文主要考虑响应变量缺失下部分线性EV模型的异方差检验问题.首先,利用完全观测到的数据对模型的未知参数和光滑函数进行估计,在此基础上利用回归借补的方法补齐缺失数据.然后,建立了对模型的随机误差进行异方差检验的经验似然比统计量,并证明该统计量渐近服从卡方分布.最后,通过数值模拟研究了检验在不同缺失概率下的有限样本性质,并在实例分析中利用部分线性EV模型对缺失数据进行了异方差检验.This paper mainly considers the problem of heteroscedasticity test for partial linear EV model with missing response variables.First,use the completely observed data to estimate the unknown parameters and smooth function of the model,and then use the regression method to fill in the missing data.Then,the empirical likelihood ratio statistic for testing the heteroscedasticity of the random error of the model is established,and it is proved that the statistic asymptotically obeys the chi-square distribution.Finally,the nature of the limited sample of the test under different missing probabilities is studied through numerical simulation,and the partial linear EV model is used in the case analysis to test the missing data for heteroscedasticity.

关 键 词：响应变量缺失 部分线性EV模型 经验似然 异方差检验 

分 类 号：O212.7[理学—概率论与数理统计]