PDF (1398K)
摘要
在半参数模型中,当数据存在非线性趋势或者模型设定错误时,其估计的稳健性结果往往不理想。传统的双重稳健估计极度依赖均值导向的假设和倾向得分模型,模型估计的稳健性得不到保障。基于线性分位数回归估计,联合优化非参数部分 g(X)与参数部分 β(τ),改进半参数分位数双重稳健估计模型,得出有效估计量。通过数值模拟假设 g(X)、β(τ)有一个及其以上被错误设定时,对比本文模型和线性分位数回归模型,模型的偏差、均方根误差均显著性低于线性分位数回归模型,且偏差预测降低了 45%左右,验证出所提模型的稳健性。在进行双重稳健估计时,采用联合优化非参数部分和参数部分,比现有的均值模型稳健性更高,更适合用于模型错误设定或存在非线性趋势的模型中。
Abstract
In semi-parametric models, estimation robustness is often compromised by nonlinear trends or model misspecification. Traditional doubly robust estimators, which rely heavily on mean-oriented assumptions and propensity score models, frequently fail to deliver reliable inference. This paper develops an enhanced semi-parametric quantile doubly robust estimation model that integrates local linear quantile regression to simultaneously optimize both the nonparametric component g(X) and the parametric component β(τ), thereby obtaining efficient estimators. Numerical simulations-conducted under scenarios where at least one of g(X) or β(τ) is misspecified-show that the proposed model achieves significantly lower bias and root mean square error compared to conventional linear quantile regression. Specifically, bias is reduced by approximately 45%, highlighting the model's superior robustness. By adopting a joint optimization strategy for both components, the proposed framework exhibits stronger robustness than existing mean-based models, making it particularly suitable for applications involving model misspecification or nonlinear trends.
关键词
Key words
[Author(id=1330580023703057108, tenantId=1045748351789510663, journalId=1155139928303341748, articleId=1330569258212086685, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=zy011405@163.com, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1330580023761777368, tenantId=1045748351789510663, journalId=1155139928303341748, articleId=1330569258212086685, authorId=1330580023703057108, language=EN, stringName=Yun ZENG, firstName=Yun, middleName=null, lastName=ZENG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Mathematics and Statistics, Chongqing Technology and Business University , Chongqing 400067, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1330580023812109018, tenantId=1045748351789510663, journalId=1155139928303341748, articleId=1330569258212086685, authorId=1330580023703057108, language=CN, stringName=曾赟, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=重庆工商大学 数学与统计学院 , 重庆市 400067, bio={"content":"曾赟,硕士生;研究方向:统计推断; zy011405@163.com
"}, bioImg=null, bioContent=曾赟,硕士生;研究方向:统计推断; zy011405@163.com
, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1330580023623365326, tenantId=1045748351789510663, journalId=1155139928303341748, articleId=1330569258212086685, xref=null, ext=[AuthorCompanyExt(id=1330580023640142543, tenantId=1045748351789510663, journalId=1155139928303341748, articleId=1330569258212086685, companyId=1330580023623365326, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Mathematics and Statistics, Chongqing Technology and Business University , Chongqing 400067, China), AuthorCompanyExt(id=1330580023652725456, tenantId=1045748351789510663, journalId=1155139928303341748, articleId=1330569258212086685, companyId=1330580023623365326, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=重庆工商大学 数学与统计学院 , 重庆市 400067)])])]
曾赟.
半参数模型稳健性的改进---基于分位数双重稳健估计[J].
齐鲁工业大学学报, 2026, 40(4): 71-80 DOI:10.16442/j.cnki.qlgydxxb.2026.04.008
| [1] |
Koenker R, Bassett G J. Regression quantiles[J]. Econometrica, 1978, 46(1): 33-50.
|
| [2] |
de Boor C. A practical guide to splines[M]. New York: Springer, 1978.
|
| [3] |
Hastie T, Tibshirani R. Generalized additive models: some applications[J]. Journal of the American Statistical Association, 1987, 82(398): 371-386.
|
| [4] |
Robins J M, Rotnitzky A. Semiparametric efficiency in multivariate regression models with missing data[J]. Journal of the American Statistical Association, 1995, 90(429): 122-129.
|
| [5] |
Laan M J, Robins J M. Unified methods for censored longitudinal data and causality[M]. New York: Springer, 2003.
|
| [6] |
van der Vaart A W. Asymptotic statistics[M]. Cambridge: Cambridge University Press, 2000.
|
| [7] |
Portnoy S. Local asymptotics for quantile smoothing splines[J]. The Annals of Statistics, 1997, 25(1): 414-434.
|
| [8] |
Koenker R, Ng P, Portnoy S. Quantile smoothing splines[J]. Biometrika, 1994, 81(4): 673-680.
|
| [9] |
Stone C J. Optimal global rates of convergence for nonparametric regression[J]. The Annals of Statistics, 1982, 10(4): 1040-1053.
|
| [10] |
Bang H, Robins J M. Doubly robust estimation in missing data and causal inference models[J]. Biometrics, 2005, 61(4): 962-973.
|
| [11] |
Koenker R. Quantile regression[M]. Cambridge: Cambridge University Press, 2005.
|
| [12] |
Zou Hui, Yuan Ming. Composite quantile regression and the oracle model selection theory[J]. The Annals of Statistics, 2008, 36(3): 1108-1126.
|
| [13] |
Tan Zhiqiang. Bounded, efficient and doubly robust estimation with inverse weighting[J]. Biometrika, 2010, 97(3): 661-682.
|
| [14] |
Zheng Qi, Peng Limin, He Xuming. Globally adaptive quantile regression with ultra-high dimensional data[J]. The Annals of Statistics, 2015, 43(5): 2225-2258.
|
| [15] |
Chernozhukov V, Hansen C. Instrumental variable quantile regression: A robust inference approach[J]. Journal of Econometrics, 2008, 142(1): 379-398.
|
| [16] |
Belloni A, Chernozhukov V, Chetverikov D, et al. Uniformly valid post-regularization confidence regions for many functional parameters in z-estimation framework[J]. The Annals of Statistics, 2018, 46(6B): 3643-3675.
|
| [17] |
Caracciolo F, Furno M. Quantile treatment effect and double robust estimators: an appraisal on the Italian labor market[J]. Journal of Economic Studies, 2017, 44(4): 585-604.
|
| [18] |
Liu Mollei, Zhang Yi, Zhou Doudou. Double/debiased machine learning for logistic partially linear model[J]. The Econometrics Journal, 2021, 24(3): 559-588.
|
| [19] |
丁先文, 张文, 袁红, 等. 含缺失数据的半参数模型的稳健估计[J]. 统计与决策, 2022, 38(1): 25-28.
|
| [20] |
田玲, 孙宁, 杨琛. 基于半参数分位数回归法的地震损失研究[J]. 统计与决策, 2019, 35(9): 11-14.
|
| [21] |
许文超, 张日权, 梁华. 两样本广义函数型线性模型的估计和半参数有效性[J]. 中国科学:数学, 2025, 55(6): 1123-1158.
|
| [22] |
张日权. 现代非参数统计方法[M]. 北京: 科学出版社, 2012.
|
| [23] |
Chernozhukov V, Escanciano J C, Ichimura H, et al. Locally robust semiparametric estimation[J]. Econometrica, 2022, 90(4): 1501-1535.
|
| [24] |
Blanchet J, Li Jiajin, Lin Sirui, et al. Distributionally robust optimization and robust statistics[J]. Statistical Science, 2025, 40(3): 351-377.
|
基金资助
重庆市自然科学基金项目(CSTB2022NSCQ-MSX1370)