A New Basis for Sparse Principal Component Analysis

Fan Chen, Karl Rohe

Previous versions of sparse principal component analysis (PCA) have presumed that the eigen-basis (a $p\times k$ matrix) is approximately sparse. We propose a method that presumes…