Optimal bandwidths for kernel density estimators of functions of observations

    Authors

    I. A. Ahmad;Y. Q. Fan

    Abbreviated Journal Title

    Stat. Probab. Lett.

    Keywords

    density estimation; kernel methods; functions of random variables; optimal bandwidths; integrated mean square error; SELECTION; DERIVATIVES; Statistics & Probability

    Abstract

    We derive optimal bandwidths for kernel density estimators of functions of observations proposed in Frees (J. Amer. Statist. Assoc. 89 (1994) 517-525). Our criteria are, respectively, the minimization of the asymptotic mean squared error and of the asymptotic mean integrated squared error of the estimators. (C) 2001 Elsevier Science B.V. All rights reserved.

    Journal Title

    Statistics & Probability Letters

    Volume

    51

    Issue/Number

    3

    Publication Date

    1-1-2001

    Document Type

    Article

    Language

    English

    First Page

    245

    Last Page

    251

    WOS Identifier

    WOS:000166704500005

    ISSN

    0167-7152

    Share

    COinS