Estimating quantiles of a population is a fundamental problem in nonparametric statistics, with high practical relevance. This note deals, from a Bayesian point of view, with quantile estimation in deconvolution problems with known error distribution. We pursue the analysis for error distributions whose characteristic functions decay polynomially fast, the so-called ordinary smooth errors that lead to mildly ill-posed inverse problems. Our method is based on Fourier inversion techniques for density deconvolution and the estimation procedure for single quantiles is minimax-optimal under minimal conditions.
Bayesian quantile estimation in deconvolution
Catia Scricciolo
2021-01-01
Abstract
Estimating quantiles of a population is a fundamental problem in nonparametric statistics, with high practical relevance. This note deals, from a Bayesian point of view, with quantile estimation in deconvolution problems with known error distribution. We pursue the analysis for error distributions whose characteristic functions decay polynomially fast, the so-called ordinary smooth errors that lead to mildly ill-posed inverse problems. Our method is based on Fourier inversion techniques for density deconvolution and the estimation procedure for single quantiles is minimax-optimal under minimal conditions.File in questo prodotto:
Non ci sono file associati a questo prodotto.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.