High-dimensional quantile regression: Convolution smoothing and concave regularization

Kean Ming Tan, Lan Wang, Wen Xin Zhou

Research output: Contribution to journalArticlepeer-review


(Formula presented.) -penalized quantile regression (QR) is widely used for analysing high-dimensional data with heterogeneity. It is now recognized that the (Formula presented.) -penalty introduces non-negligible estimation bias, while a proper use of concave regularization may lead to estimators with refined convergence rates and oracle properties as the signal strengthens. Although folded concave penalized M-estimation with strongly convex loss functions have been well studied, the extant literature on QR is relatively silent. The main difficulty is that the quantile loss is piecewise linear: it is non-smooth and has curvature concentrated at a single point. To overcome the lack of smoothness and strong convexity, we propose and study a convolution-type smoothed QR with iteratively reweighted (Formula presented.) -regularization. The resulting smoothed empirical loss is twice continuously differentiable and (provably) locally strongly convex with high probability. We show that the iteratively reweighted (Formula presented.) -penalized smoothed QR estimator, after a few iterations, achieves the optimal rate of convergence, and moreover, the oracle rate and the strong oracle property under an almost necessary and sufficient minimum signal strength condition. Extensive numerical studies corroborate our theoretical results.

Original languageEnglish (US)
Pages (from-to)205-233
Number of pages29
JournalJournal of the Royal Statistical Society. Series B: Statistical Methodology
Issue number1
StatePublished - Feb 2022
Externally publishedYes


  • concave regularization
  • convolution
  • minimum signal strength
  • oracle property
  • quantile regression

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty


Dive into the research topics of 'High-dimensional quantile regression: Convolution smoothing and concave regularization'. Together they form a unique fingerprint.

Cite this