Nonparametric estimation of multivariate scale mixtures of uniform densities

Nonparametric estimation (62G05) Density estimation (62G07) Asymptotic properties of nonparametric inference (62G20) Estimation in multivariate analysis (62H12)

Abstract: Suppose that

m U = (U_{1}, l d o t s, U_{d})

has a Uniform

([0, 1]^{d})

distribution, that

m Y = (Y_{1}, l d o t s, Y_{d})

has the distribution

G

on

R R_{+}^{d}

, and let

m X = (X_{1}, l d o t s, X_{d}) = (U_{1} Y_{1}, l d o t s, U_{d} Y_{d})

. The resulting class of distributions of

m X

(as

G

varies over all distributions on

R R_{+}^{d}

) is called the {sl Scale Mixture of Uniforms} class of distributions, and the corresponding class of densities on

R R_{+}^{d}

is denoted by

{c a l F}_{S M U} (d)

. We study maximum likelihood estimation in the family

{c a l F}_{S M U} (d)

. We prove existence of the MLE, establish Fenchel characterizations, and prove strong consistency of the almost surely unique maximum likelihood estimator (MLE) in

{c a l F}_{S M U} (d)

. We also provide an asymptotic minimax lower bound for estimating the functional

f m a p s t o f (m x)

under reasonable differentiability assumptions on

f i n {c a l F}_{S M U} (d)

in a neighborhood of

m x

. We conclude the paper with discussion, conjectures and open problems pertaining to global and local rates of convergence of the MLE.

Recommendations

Cites work

Cited in

(15)

This page was built for publication: Nonparametric estimation of multivariate scale mixtures of uniform densities

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q413749)