Linear functional regression: the case of fixed design and functional response

被引：115

作者：

Cuevas, A ^{[1
]}

Febrero, M ^{[1
]}

Fraiman, R ^{[1
]}

机构：

[1] Univ Autonoma Madrid, Dept Matemat, E-28049 Madrid, Spain

来源：

CANADIAN JOURNAL OF STATISTICS-REVUE CANADIENNE DE STATISTIQUE | 2002年 / 30卷 / 02期

关键词：

calibration; consistent estimation; functional data; functional regression; linear operators; Yurinskii's inequality;

D O I：

10.2307/3315952

中图分类号：

O21 [概率论与数理统计]; C8 [统计学];

学科分类号：

020208 ; 070103 ; 0714 ;

摘要：

The authors consider the problem of simple linear regression when the exogenous and endogenous variables are functional and the design is fixed. They propose an estimator for the underlying linear operator and prove its consistency under some conditions which ensure that the design is sufficiently informative. They consider the classical calibration (or inverse regression) problem and analyze a consistent estimator. They also give a simulation study. The proposed method is not hard to implement in practice.

引用

页码：285 / 300

页数：16