The Use of Bootstrap in System Identification
1999 (English)Report (Other academic)
Simulation based methods used to estimate uncertainties in models have recently gained much interest. We study the bootstrap and how this method could be used in system identification related problems. The main contribution is a proposed algorithm to estimate the probability density function in case of undermodeling. We illustrate the performance of the bootstrap resampling method by simulation examples, which are in good agreement with Monte Carlo simulations.
Place, publisher, year, edition, pages
Linköping: Linköping University Electronic Press, 1999. , 20 p.
LiTH-ISY-R, ISSN 1400-3902 ; 2113
Simulation, Estimation, Models, Identification, Performance, Bootstrap
IdentifiersURN: urn:nbn:se:liu:diva-55696ISRN: LiTH-ISY-R-2113OAI: oai:DiVA.org:liu-55696DiVA: diva2:316454