Aportes a la Teoría y la Implementación del Método LSCR
2010 (Spanish)In: RIAI - Revista Iberoamericana de Automatica e Informatica Industrial, ISSN 1697-7912, Vol. 7, no 3, 83-94 p.Article in journal (Refereed) Published
The LSCR method (Leave-out-Sign-dominant-Correlation-Regions) provides confidence regions for the parameters of a system by evaluating a set of correlation functions calculated for the available data. To do the approximation for the whole region, the procedure must be repeated for each value of the parameter vector. The main attributes of LSCR are its validity for a finite set of data and the scarce assumptions on the noise. However, the procedure needs much computational effort, which limits its application to very simple cases. In this work some theoretical aspects of the LSCR method are improved and some implementation alternatives are suggested. It is also compared, in terms of computational effort, with Bootstrap, another way to obtain confidence regions.
Place, publisher, year, edition, pages
2010. Vol. 7, no 3, 83-94 p.
Modeling error, Prediction error, Parameter estimation, Uncertainty
Computer and Information Science
IdentifiersURN: urn:nbn:se:kth:diva-29413DOI: 10.4995/RIAI.2010.03.08ISI: 000280581500008ScopusID: 2-s2.0-78650346976OAI: oai:DiVA.org:kth-29413DiVA: diva2:398147
QC 201102162011-02-162011-02-022016-06-01Bibliographically approved