Title :
A parametric version of jackknife-after-bootstrap
Author_Institution :
Dept. of Math. & Comput. Sci., Valdosta State Univ., GA, USA
Abstract :
We investigate the problem of deriving precision estimates for bootstrap quantities within parametric families. B. Efron´s (1992) jackknife-after-bootstrap is a simple approach that only uses the information in the original bootstrap samples via the importance sampling technique, with no further resampling required. This method can be applied to many Monte Carlo experiments, especially, to the parametric input modeling problems. Variance analysis of the parametric jackknife-after-bootstrap is discussed. Under some reasonable conditions, the parametric jackknife-after-bootstrap method is as good as the true jackknife method. A generalized parametric jackknife-after-bootstrap method is introduced
Keywords :
importance sampling; parameter estimation; simulation; Monte Carlo experiments; bootstrap quantities; bootstrap samples; generalized parametric method; importance sampling technique; jackknife method; jackknife-after-bootstrap method; parametric families; parametric input modeling problems; parametric version; precision estimates; variance analysis; Analysis of variance; Application software; Computational modeling; Computer errors; Computer science; Mathematical model; Mathematics; Monte Carlo methods; Parametric statistics; Power system modeling;
Conference_Titel :
Simulation Conference Proceedings, 1998. Winter
Conference_Location :
Washington, DC
Print_ISBN :
0-7803-5133-9
DOI :
10.1109/WSC.1998.745038