An iterative coordinate descent algorithm to compute sparse low-rank approximations
This work addresses the need for efficient sparse principal component analysis in machine learning and data science, though it appears incremental as an extension of existing algorithms.
The paper tackles the problem of computing sparse low-rank approximations for data matrices by proposing a new iterative coordinate descent algorithm that extends the Kogbetliantz method to build approximate singular value decompositions for a few principal components, and it demonstrates performance in recovering sparse principal components on various datasets and in dimensionality reduction for classification.
In this paper, we describe a new algorithm to build a few sparse principal components from a given data matrix. Our approach does not explicitly create the covariance matrix of the data and can be viewed as an extension of the Kogbetliantz algorithm to build an approximate singular value decomposition for a few principal components. We show the performance of the proposed algorithm to recover sparse principal components on various datasets from the literature and perform dimensionality reduction for classification applications.