Bootstrap prediction intervals for multivariate time series
Keywords and Phrases
"The theory and methodology of obtaining bootstrap prediction intervals for univariate time series using the forward representation of the series is extended to vector autoregressive (VAR) models. Kim has shown that simultaneous prediction intervals based on the Bonferroni method and the backward representation of the time series achieve coverage close to nominal when the parameter estimates are corrected for small sample bias. To utilize his method, it is necessary to assume that the innovations are normally distributed to maintain independence of the innovations associated with the backward representation of the time series. This assumption is not necessary if the forward representation is used. Bootstrap prediction intervals based on the forward representation of the time series, are less restrictive and thus can also be adapted for time series that do not have a backward representation. The asymptotic validity of the proposed bootstrap method is established and small sample properties are studied using Monte Carlo simulation"--Abstract, leaf iii.
Mathematics and Statistics
Ph. D. in Mathematics
University of Missouri--Rolla
x, 170 leaves
© 2005 Florian Sebastian Rueck, All rights reserved.
Dissertation - Citation
Library of Congress Subject Headings
Print OCLC #
Link to Catalog Record
Full-text not available: Request this publication directly from Missouri S&T Library or contact your local library.http://laurel.lso.missouri.edu/record=b5595070~S5
Rueck, Florian Sebastian, "Bootstrap prediction intervals for multivariate time series" (2005). Doctoral Dissertations. 1621.
This document is currently not available here.