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

Statistics > Methodology

Title: A Nonparametric Bayesian Technique for High-Dimensional Regression

Abstract: This paper proposes a nonparametric Bayesian framework called VariScan for simultaneous clustering, variable selection, and prediction in high-throughput regression settings. Poisson-Dirichlet processes are utilized to detect lower-dimensional latent clusters of covariates. An adaptive nonlinear prediction model is constructed for the response, achieving a balance between model parsimony and flexibility. Contrary to conventional belief, cluster detection is shown to be aposteriori consistent for a general class of models as the number of covariates and subjects grows. Simulation studies and data analyses demonstrate that VariScan often outperforms several well-known statistical methods.
Comments: arXiv admin note: substantial text overlap with arXiv:1407.5472
Subjects: Methodology (stat.ME)
DOI: 10.1214/16-EJS1184
Cite as: arXiv:1604.03615 [stat.ME]
  (or arXiv:1604.03615v1 [stat.ME] for this version)

Submission history

From: Subharup Guha [view email]
[v1] Tue, 12 Apr 2016 23:24:00 GMT (401kb,D)

Link back to: arXiv, form interface, contact.