English

Adaptive thresholding estimation of a Poisson intensity with infinite support

Statistics Theory 2008-01-22 v1 Statistics Theory

Abstract

The purpose of this paper is to estimate the intensity of a Poisson process NN by using thresholding rules. In this paper, the intensity, defined as the derivative of the mean measure of NN with respect to ndxndx where nn is a fixed parameter, is assumed to be non-compactly supported. The estimator f~n,γ\tilde{f}_{n,\gamma} based on random thresholds is proved to achieve the same performance as the oracle estimator up to a logarithmic term. Oracle inequalities allow to derive the maxiset of f~n,γ\tilde{f}_{n,\gamma}. Then, minimax properties of f~n,γ\tilde{f}_{n,\gamma} are established. We first prove that the rate of this estimator on Besov spaces Bp,q\al{\cal B}^\al_{p,q} when p2p\leq 2 is (ln(n)/n)\al/(1+2\al)(\ln(n)/n)^{\al/(1+2\al)}. This result has two consequences. First, it establishes that the minimax rate of Besov spaces Bp,q\al{\cal B}^\al_{p,q} with p2p\leq 2 when non compactly supported functions are considered is the same as for compactly supported functions up to a logarithmic term. This result is new. Furthermore, f~n,γ\tilde{f}_{n,\gamma} is adaptive minimax up to a logarithmic term. When p>2p>2, the situation changes dramatically and the rate of f~n,γ\tilde{f}_{n,\gamma} on Besov spaces Bp,q\al{\cal B}^\al_{p,q} is worse than (ln(n)/n)\al/(1+2\al)(\ln(n)/n)^{\al/(1+2\al)}. Finally, the random threshold depends on a parameter γ\gamma that has to be suitably chosen in practice. Some theoretical results provide upper and lower bounds of γ\gamma to obtain satisfying oracle inequalities. Simulations reinforce these results.

Keywords

Cite

@article{arxiv.0801.3157,
  title  = {Adaptive thresholding estimation of a Poisson intensity with infinite support},
  author = {Patricia Reynaud-Bouret and Vincent Rivoirard},
  journal= {arXiv preprint arXiv:0801.3157},
  year   = {2008}
}
R2 v1 2026-06-21T10:04:49.233Z