English

Random threshold for linear model selection, revisited

Methodology 2010-10-27 v1

Abstract

In [Lavielle and Ludena 07], a random thresholding metho d is intro duced to select the significant, or non null, mean terms among a collection of independent random variables, and applied to the problem of recovering the significant coefficients in non ordered model selection. We intro duce a simple modification which removes the dep endency of the proposed estimator on a window parameter while maintaining its asymptotic properties. A simulation study suggests that both procedures compare favorably to standard thresholding approaches, such as multiple testing or model-based clustering, in terms of the binary classification risk. An application of the method to the problem of activation detection on functional magnetic resonance imaging (fMRI) data is discussed.

Keywords

Cite

@article{arxiv.1010.5389,
  title  = {Random threshold for linear model selection, revisited},
  author = {Merlin Keller and Marc Lavielle},
  journal= {arXiv preprint arXiv:1010.5389},
  year   = {2010}
}

Comments

22 pages, 7 figures. Submitted to Statistics and its Interface (SII)

R2 v1 2026-06-21T16:34:16.663Z