复杂纵向数据变系数模型的估计和变量选择研究新进展  

New Progress in Estimation and Variable Selection of Variable Coefficient Models for Complex Longitudinal Data

作　　者：付学宝 赵明涛[1] 贺大伟 Fu Xuebao;Zhao Mingtao;He Dawei(School of Statistics and Applied Mathematics,Anhui University of Finance and Economics,Bengbu Anhui 233030,China;School of Arts and Letters,Anhui University of Finance and Economics,Bengbu Anhui 233030,China)

机构地区：[1]安徽财经大学统计与应用数学学院,安徽蚌埠233030 [2]安徽财经大学文学院,安徽蚌埠233030

出　　处：《统计与决策》2025年第5期68-73,共6页Statistics & Decision

基　　金：安徽省哲学社会科学一般项目(AHSKF2022D08)。

摘　　要：变系数模型是一类重要的结构化半参数模型,常被用于复杂数据的半参数统计推断,其模型估计和变量选择问题是变系数模型研究的重要内容。文章系统介绍了复杂纵向数据变系数模型估计和变量选择的研究新进展,包括变系数模型、纵向数据变系数模型、纵向数据变系数EV模型、纵向缺失数据变系数EV模型。最后,阐述了在方法改进、理论研究与模型验证方面尚待解决的一些问题,并对未来该领域的研究进行了展望。Variable coefficient model is an important structured semi-parametric model,often used for semi-parametric statistical inference of complex data.The issues of model estimation and variable selection are important elements in the study of variable coefficient models.This paper introduces the new developments in the estimation and variable selection of complex longitudinal data variable coefficient models,including the variable coefficient model,the longitudinal data variable coefficient model,the longitudinal data variable coefficient EV model,and the longitudinal missing data variable coefficient EV model.Finally,the paper points out some problems remaining to be solved in method improvement,theoretical research,and model validation,and explores the prospects for future researches.

关 键 词：复杂纵向数据 变系数模型 变系数EV模型 二次推断函数 

分 类 号：O212[理学—概率论与数理统计]