Resampling m-dependent random variables with applications to forecasting

Authors
Citation
S. Sjostedt, Resampling m-dependent random variables with applications to forecasting, SC J STAT, 27(3), 2000, pp. 543-561
Citations number
23
Categorie Soggetti
Mathematics
Journal title
SCANDINAVIAN JOURNAL OF STATISTICS
ISSN journal
03036898 → ACNP
Volume
27
Issue
3
Year of publication
2000
Pages
543 - 561
Database
ISI
SICI code
0303-6898(200009)27:3<543:RMRVWA>2.0.ZU;2-G
Abstract
Resampling methods are proposed to estimate the distributions of sums of m- dependent possibly differently distributed real-valued random variables. Th e random variables are allowed to have varying mean values. A non-parametri c resampling method based on the moving blocks bootstrap is proposed for th e case in which the mean values are smoothly varying or "asymptotically equ al". The idea Is to resample blocks in pairs, It is also confirmed that a " circular" block resampling scheme can be used in the case where the mean va lues are "asymptotically equal". A central limit resampling theorem for eac h of the two cases is proved. The resampling methods have a potential appli cation to time series analysis, to distinguish between two different foreca sting models. This is illustrated with an example using Swedish export pric es of coated paper products.