ON THE SPECTRAL PROPERTIES OF MATRICES ASSOCIATED WITH TREND FILTERS

被引：8

作者：

Luati, Alessandra ^{[1
]}

Proietti, Tommaso ^{[2
]}

机构：

[1] Univ Bologna, Dept Stat, I-40126 Bologna, Italy

[2] Univ Roma Tor Vergata, Rome, Italy

来源：

ECONOMETRIC THEORY | 2010年 / 26卷 / 04期

关键词：

BUSINESS CYCLES; TOEPLITZ;

D O I：

10.1017/S0266466609990715

中图分类号：

F [经济];

学科分类号：

02 ;

摘要：

This note is concerned with the spectral properties of matrices associated with linear smoothers. We derive analytical results on the eigenvalues and eigenvectors of smoothing matrices by interpreting the latter as perturbations of matrices belonging to algebras with known spectral properties, such as the circulant and the generalized tau. These results are used to characterize the properties of a smoother in terms of an approximate eigen-decomposition of the associated smoothing matrix.

引用

页码：1247 / 1261

页数：15

共 39 条

[1]

[Anonymous], 1996, Local polynomial modelling and its applications

[2]

[Anonymous], 2003, Semiparametric Regression

[3]

[Anonymous], 2002, Int. J. Math. Educ. Sci. Technol, DOI DOI 10.1080/00207390110118953

[4]

[Anonymous], 2022, locfit: Local regression, likelihood and density estimation