BOOTSTRAP CONFIDENCE INTERVAL FOR MEDIAN

Suprihatin, Bambang and Guritno, Suryo and Haryatmi, Sri (2013) BOOTSTRAP CONFIDENCE INTERVAL FOR MEDIAN. Scientific Conference, 1. ISSN 978-3-937535-89-3

[img] Microsoft Word (DCM Verlag Berlin) - Published Version
Download (334Kb)

    Abstract

    Let sample at hand and let X of size n from an unknown distribution F. If we consider that all elements of X are distinct, then the number of different possible resamples with replacement equals to . In general, this number obvious very large for larger n.. For instance, if n = 10, then the number is an enormous number. Let be the estimate value of statistic computed from , where t is functional. In most cases of practical interest, each distinct (without regard for order), gives rise to a distinct . Accordingly, we concern only on so-called atoms of nonparametric bootstrap. The number of atoms is far less than Based on these atoms, the nonparametric bootstrap used to estimate a statistic computed from X. This paper presents how to find the number of atoms. The implementation of the uses of atoms is applied in bootstrapping bias estimate of sample median. Bootstrap version of standar error as a measure of accuracy of estimator is considered, as well. The main purpose of this paper is to construct a confidence interval for median. Results from Monte Carlo simulation for these cases are also presented.

    Item Type: Article
    Subjects: Q Science > QA Mathematics
    Divisions: Faculty of Mathematics and Natural Sciences > Department of Mathematics
    Depositing User: Mr. Bambang Suprihatin
    Date Deposited: 22 Apr 2014 18:38
    Last Modified: 22 Apr 2014 18:38
    URI: http://eprints.unsri.ac.id/id/eprint/4337

    Actions (login required)

    View Item