English

Penalized contrast estimator for adaptive density deconvolution

Statistics Theory 2008-02-11 v1 Statistics Theory

Abstract

The authors consider the problem of estimating the density gg of independent and identically distributed variables X_iX\_i, from a sample Z_1,...,Z_nZ\_1, ..., Z\_n where Z_i=X_i+σϵ_iZ\_i=X\_i+\sigma\epsilon\_i, i=1,...,ni=1, ..., n, ϵ\epsilon is a noise independent of XX, with σϵ\sigma\epsilon having known distribution. They present a model selection procedure allowing to construct an adaptive estimator of gg and to find non-asymptotic bounds for its L_2(R)\mathbb{L}\_2(\mathbb{R})-risk. The estimator achieves the minimax rate of convergence, in most cases where lowers bounds are available. A simulation study gives an illustration of the good practical performances of the method.

Keywords

Cite

@article{arxiv.math/0601091,
  title  = {Penalized contrast estimator for adaptive density deconvolution},
  author = {Fabienne Comte and Yves Rozenholc and Marie-Luce Taupin},
  journal= {arXiv preprint arXiv:math/0601091},
  year   = {2008}
}