An error bound for L1-norm support vector machine coefficients in ultra-high dimension

Bo Peng, Lan Wang, Yichao Wu

Research output: Contribution to journalArticlepeer-review

24 Scopus citations


Comparing with the standard L2-norm support vector machine (SVM), the L1-norm SVM enjoys the nice property of simultaneously preforming classification and feature selection. In this paper, we investigate the statistical performance of L1-norm SVM in ultra-high dimension, where the number of features p grows at an exponential rate of the sample size n. Different from existing theory for SVM which has been mainly focused on the generalization error rates and empirical risk, we study the asymptotic behavior of the coefficients of L1-norm SVM. Our analysis reveals that the estimated L1-norm SVM coefficients achieve near oracle rate, that is, with high probability, the L2 error bound of the estimated L1-norm SVM coefficients is of order Op(√q log p/n), where q is the number of features with nonzero coefficients. Furthermore, we show that if the L1-norm SVM is used as an initial value for a recently proposed algorithm for solving non-convex penalized SVM (Zhang et al., 2016b), then in two iterative steps it is guaranteed to produce an estimator that possesses the oracle property in ultra-high dimension, which in particular implies that with probability approaching one the zero coefficients are estimated as exactly zero. Simulation studies demonstrate the fine performance of L1-norm SVM as a sparse classifier and its effectiveness to be utilized to solve non-convex penalized SVM problems in high dimension.

Original languageEnglish (US)
Pages (from-to)1-26
Number of pages26
JournalJournal of Machine Learning Research
StatePublished - Dec 1 2016

Bibliographical note

Funding Information:
We would like to acknowledge support for this project from the National Science Foundation (NSF grant DMS-1308960) for Peng and Wang's research and the National Science Foundation (NSF grant DMS-1055210) and National Institutes of Health (NIH grants R01 CA149569 and P01 CA142538) for Wu's research respectively.

Publisher Copyright:
© 2016 Bo Peng, Lan Wang and Yichao Wu.


  • Error bound
  • Feature selection
  • L-norm SVM
  • Non-convex penalty
  • Oracle property
  • Support vector machine
  • Ulta-high dimension


Dive into the research topics of 'An error bound for L1-norm support vector machine coefficients in ultra-high dimension'. Together they form a unique fingerprint.

Cite this