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.
Publié le : 1992-03-14
Classification:
Bootstrap,
time series,
kernel spectral estimate,
periodogram,
bandwidth selection,
62M15,
62G05
@article{1176348515,
author = {Franke, J. and Hardle, W.},
title = {On Bootstrapping Kernel Spectral Estimates},
journal = {Ann. Statist.},
volume = {20},
number = {1},
year = {1992},
pages = { 121-145},
language = {en},
url = {http://dml.mathdoc.fr/item/1176348515}
}
Franke, J.; Hardle, W. On Bootstrapping Kernel Spectral Estimates. Ann. Statist., Tome 20 (1992) no. 1, pp. 121-145. http://gdmltest.u-ga.fr/item/1176348515/