A nearest neighbor estimate of the residual variance

We study the problem of estimating the smallest achievable mean-squared error in regression function estimation. The problem is equivalent to estimating the second moment of the regression function of Y on X∈Rd. We introduce a nearest-neighbor-based estimate and obtain a normal limit law for the est...

ver descrição completa

Detalhes bibliográficos
Autores: Devroye, Luc, Györfi, László, Lugosi, Gábor, Walk, Harro
Tipo de documento: artigo
Estado:Versão publicada
Data de publicação:2018
País:España
Recursos:Universitat Pompeu Fabra
Repositório:Repositorio Digital de la UPF
OAI Identifier:oai:repositori.upf.edu:10230/44868
Acesso em linha:http://hdl.handle.net/10230/44868
http://dx.doi.org/10.1214/18-EJS1438
Access Level:Acceso aberto
Palavra-chave:Regression functional
Nearest-neighbor-based estimate
Asymptotic normality
Concentration inequalities
Dimension reduction
Descrição
Resumo:We study the problem of estimating the smallest achievable mean-squared error in regression function estimation. The problem is equivalent to estimating the second moment of the regression function of Y on X∈Rd. We introduce a nearest-neighbor-based estimate and obtain a normal limit law for the estimate when X has an absolutely continuous distribution, without any condition on the density. We also compute the asymptotic variance explicitly and derive a non-asymptotic bound on the variance that does not depend on the dimension d. The asymptotic variance does not depend on the smoothness of the density of X or of the regression function. A non-asymptotic exponential concentration inequality is also proved. We illustrate the use of the new estimate through testing whether a component of the vector X carries information for predicting Y.