基于FPCA的部分函数型线性模型的复合分位数回归估计  

Composite Quantile Regression for Partial Functional Linear Regression Model based on FPCA

作　　者：余平[1] YU Ping(College of Mathematics and Computer Science,Shanxi Normal University,Linfen 041000,Shanxi,China)

机构地区：[1]山西师范大学数学与计算机科学学院

出　　处：《山西师范大学学报（自然科学版）》2019年第3期5-12,共8页Journal of Shanxi Normal University(Natural Science Edition)

摘　　要：本文研究了部分函数型线性回归模型的复合分位数估计问题.采用函数型主成分基函数对斜率函数和函数型预测变量进行展开,在相当宽松的条件下给出斜率函数的最优收敛速度和参数部分的渐近正态性.最后通过理论模拟来评价提出方法的有效性.In this paper, we propose composite quantile regression for partial functional linear model. The functional principal component analysis is employed to estimate the slope function and functional predictive variable respectively, and the convergence rates of the estimators and asymptotic normality of parametric vector are obtained under some regularity conditions. Simulation studies and real data analysis are presented for illustration of the performance of the proposed estimators.

关 键 词：部分函数型线性回归模型 复合分位数回归 函数型主成分分析 渐近正态性 

分 类 号：O212[理学—概率论与数理统计]