We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ML

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Statistics > Machine Learning

Title: MM Algorithms for Distance Covariance based Sufficient Dimension Reduction and Sufficient Variable Selection

Abstract: Sufficient dimension reduction (SDR) using distance covariance (DCOV) was recently proposed as an approach to dimension-reduction problems. Compared with other SDR methods, it is model-free without estimating link function and does not require any particular distributions on predictors (see Sheng and Yin, 2013, 2016). However, the DCOV-based SDR method involves optimizing a nonsmooth and nonconvex objective function over the Stiefel manifold. To tackle the numerical challenge, we novelly formulate the original objective function equivalently into a DC (Difference of Convex functions) program and construct an iterative algorithm based on the majorization-minimization (MM) principle. At each step of the MM algorithm, we inexactly solve the quadratic subproblem on the Stiefel manifold by taking one iteration of Riemannian Newton's method. The algorithm can also be readily extended to sufficient variable selection (SVS) using distance covariance. We establish the convergence property of the proposed algorithm under some regularity conditions. Simulation studies show our algorithm drastically improves the computation efficiency and is robust across various settings compared with the existing method. Supplemental materials for this article are available.
Comments: 26 pages, 4 figures
Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG); Computation (stat.CO)
Journal reference: Computational Statistics & Data Analysis, 2021
DOI: 10.1016/j.csda.2020.107089
Cite as: arXiv:1912.06342 [stat.ML]
  (or arXiv:1912.06342v2 [stat.ML] for this version)

Submission history

From: Runxiong Wu [view email]
[v1] Fri, 13 Dec 2019 07:08:15 GMT (342kb,D)
[v2] Wed, 3 Mar 2021 08:50:08 GMT (420kb,D)

Link back to: arXiv, form interface, contact.