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

Download:

Current browse context:

stat.ME

Change to browse by:

References & Citations

Bookmark

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

Statistics > Methodology

Title: Spike and slab Bayesian sparse principal component analysis

Authors: Bo Ning
Abstract: Sparse principal component analysis (PCA) is a popular tool for dimensional reduction of high-dimensional data. Despite its massive popularity, there is still a lack of theoretically justifiable Bayesian sparse PCA that is computationally scalable. A major challenge is choosing a suitable prior for the loadings matrix, as principal components are mutually orthogonal. We propose a spike and slab prior that meets this orthogonality constraint and show that the posterior enjoys both theoretical and computational advantages. Two computational algorithms, the PX-CAVI and the PX-EM algorithms, are developed. Both algorithms use parameter expansion to deal with the orthogonality constraint and to accelerate their convergence speeds. We found that the PX-CAVI algorithm has superior empirical performance than the PX-EM algorithm and two other penalty methods for sparse PCA. The PX-CAVI algorithm is then applied to study a lung cancer gene expression dataset. $\mathsf{R}$ package $\mathsf{VBsparsePCA}$ with an implementation of the algorithm is available on The Comprehensive R Archive Network.
Comments: 27 pages, 5 tables, 1 figures
Subjects: Methodology (stat.ME); Machine Learning (stat.ML)
MSC classes: 62C10, 62H25, 62J07
Cite as: arXiv:2102.00305 [stat.ME]
  (or arXiv:2102.00305v1 [stat.ME] for this version)

Submission history

From: Bo Ning [view email]
[v1] Sat, 30 Jan 2021 20:28:30 GMT (659kb,D)

Link back to: arXiv, form interface, contact.