Penalized composite quasi-likelihood for ultrahigh dimensional variable selection

Jelena Bradic, Jianqing Fan, Weiwei Wang

Research output: Contribution to journalArticlepeer-review

91 Scopus citations

Abstract

In high dimensional model selection problems, penalized least square approaches have been extensively used. The paper addresses the question of both robustness and efficiency of penalized model selection methods and proposes a data-driven weighted linear combination of convex loss functions, together with weighted L1-penalty. It is completely data adaptive and does not require prior knowledge of the error distribution. The weighted L1-penalty is used both to ensure the convexity of the penalty term and to ameliorate the bias that is caused by the L1-penalty. In the setting with dimensionality much larger than the sample size, we establish a strong oracle property of the method proposed that has both the model selection consistency and estimation efficiency for the true non-zero coefficients. As specific examples, we introduce a robust method of composite L1-L2, and an optimal composite quantile method and evaluate their performance in both simulated and real data examples.

Original languageEnglish (US)
Pages (from-to)325-349
Number of pages25
JournalJournal of the Royal Statistical Society. Series B: Statistical Methodology
Volume73
Issue number3
DOIs
StatePublished - Jun 2011

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Keywords

  • Composite quasi-maximum likelihood estimation
  • Lasso
  • Model selection
  • Non-polynomial dimensionality
  • Oracle property
  • Robust statistics
  • Smoothly clipped absolute deviation

Fingerprint Dive into the research topics of 'Penalized composite quasi-likelihood for ultrahigh dimensional variable selection'. Together they form a unique fingerprint.

Cite this