ON BOOTSTRAPPING KERNEL SPECTRAL ESTIMATES

被引：119

作者：

FRANKE, J ^{[1
]}

HARDLE, W ^{[1
]}

机构：

[1] CATHOLIC UNIV LOUVAIN,CORE,B-1348 LOUVAIN,BELGIUM

来源：

ANNALS OF STATISTICS | 1992年 / 20卷 / 01期

关键词：

BOOTSTRAP; TIME SERIES; KERNEL SPECTRAL ESTIMATE; PERIODOGRAM; BANDWIDTH SELECTION;

D O I：

10.1214/aos/1176348515

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：

An approach to bootstrapping kernel spectral density estimates is described which is based on resampling from the periodogram of the original data. We show that it is asymptotically valid under suitable conditions, and we illustrate its performance for a medium-sized time series sample with a small simulation study.

引用

页码：121 / 145

页数：25

共 24 条

[1]

[Anonymous], 2009, MULTIPLE TIME SERIES

[2]

BELTRAO KI, 1987, J TIME SER ANAL, V8, P21

[3] SOME ASYMPTOTIC THEORY FOR THE BOOTSTRAP [J].