Butstrapa metodes analīze atkarīgiem datiem
Loading...
Date
Authors
Advisor
Journal Title
Journal ISSN
Volume Title
Publisher
Latvijas Universitāte
Language
N/A
Abstract
Darbā ir aprakstīts butstrapa metodes pielietojums atkarīgiem datiem. Ir apskatīta bloku butstrapa metožu konstruēšana, parādīta un salīdzināta to darbības efektivitāte. Tāpat apskatīti īpaši gadījumi, kad novērtējamais parametrs ir izsakāms, kā gluda funkcija no izlases vidējās vērtības, kā kādu vienādojumu atrisinājums un kā diferencējams funkcionālis. Darba nobeigumā ir apskatīts butstrapa metodes pielietojums neparametriskās regresijas gadījumā, kad klasiskās metodes, kas balstītas uz neatkarīgiem novērojumiem, nedarbojas.
This paper is concerned with the application of bootstrap for dependent data. Block bootstrap methods are analyzed and compared to each other. Some general classes of estimators that may be represented as smooth function of sample means, as solution of equations and as differentiable functionals are reviewed. Finally, block bootstrap methods are introduced in the context of nonparametric regression where common methods are not effective or do not work.
This paper is concerned with the application of bootstrap for dependent data. Block bootstrap methods are analyzed and compared to each other. Some general classes of estimators that may be represented as smooth function of sample means, as solution of equations and as differentiable functionals are reviewed. Finally, block bootstrap methods are introduced in the context of nonparametric regression where common methods are not effective or do not work.