Approximation Algorithms for Sparse Principal Component Analysis
This work addresses the need for interpretable dimension reduction in machine learning and statistics, though it is incremental as it builds on existing sparse PCA methods.
The paper tackles the problem of improving interpretability in principal component analysis by proposing two thresholding-based approximation algorithms for sparse PCA, with the first being faster than state-of-the-art methods and the second bridging the theory-practice gap better than current approaches.
Principal component analysis (PCA) is a widely used dimension reduction technique in machine learning and multivariate statistics. To improve the interpretability of PCA, various approaches to obtain sparse principal direction loadings have been proposed, which are termed Sparse Principal Component Analysis (SPCA). In this paper, we present thresholding as a provably accurate, polynomial time, approximation algorithm for the SPCA problem, without imposing any restrictive assumptions on the input covariance matrix. Our first thresholding algorithm using the Singular Value Decomposition is conceptually simple; is faster than current state-of-the-art; and performs well in practice. On the negative side, our (novel) theoretical bounds do not accurately predict the strong practical performance of this approach. The second algorithm solves a well-known semidefinite programming relaxation and then uses a novel, two step, deterministic thresholding scheme to compute a sparse principal vector. It works very well in practice and, remarkably, this solid practical performance is accurately predicted by our theoretical bounds, which bridge the theory-practice gap better than current state-of-the-art.