Wild bootstrap

Michael Martin

    Research output: Chapter in Book/Report/Conference proceedingEntry for encyclopedia/dictionarypeer-review

    Abstract

    The wild bootstrap is a resampling method designed for regression problems with heteroscedastic error structure. The name derives from its use of a single residual to estimate the entire conditional distribution of each response through the use of an auxiliary zero-mean, unit variance random multiplier when resampling. This article describes the development, properties, and software implementation of the wild bootstrap and some of its extensions.
    Original languageEnglish
    Title of host publicationWiley StatsRef: Statistics Reference Online
    EditorsWiley StatsRef
    Place of Publicationonline
    PublisherJohn Wiley & Sons Ltd.
    Pages1-6pp
    Volume8
    DOIs
    Publication statusPublished - 2017

    Fingerprint

    Dive into the research topics of 'Wild bootstrap'. Together they form a unique fingerprint.

    Cite this