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


Current browse context:


Change to browse by:

References & Citations


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

Mathematics > Statistics Theory

Title: PAC-Bayes Bounds for High-Dimensional Multi-Index Models with Unknown Active Dimension

Abstract: The multi-index model with sparse dimension reduction matrix is a popular approach to circumvent the curse of dimensionality in a high-dimensional regression setting. Building on the single-index analysis by Alquier, P. & Biau, G. (Journal of Machine Learning Research 14 (2013) 243-280), we develop a PAC-Bayesian estimation method for a possibly misspecified multi-index model with unknown active dimension and an orthogonal dimension reduction matrix. Our main result is a non-asymptotic oracle inequality, which shows that the estimation method adapts to the active dimension of the model, the sparsity of the dimension reduction matrix and the regularity of the link function. Under a Sobolev regularity assumption on the link function the estimator achieves the minimax rate of convergence (up to a logarithmic factor) and no additional price is paid for the unknown active dimension.
Subjects: Statistics Theory (math.ST)
Cite as: arXiv:2303.13474 [math.ST]
  (or arXiv:2303.13474v1 [math.ST] for this version)

Submission history

From: Maximilian F. Steffen [view email]
[v1] Thu, 23 Mar 2023 17:45:46 GMT (18kb)

Link back to: arXiv, form interface, contact.