A study of the optimality of PCA under spectral sparsification

Abstract

Principal component analisys (PCA) is a data analysis technique for mapping points in Rn to a two or three dimensional space. This dimensionality reduction preserves the natural grouping of points and information of data.