Output details

Generalized power method for sparse principal component analysis

This gives a most competitive yet simple algorithm for the problem of sparse PCA, reformulating the optimization problem as maximization of a convex function on a compact convex set. One of the first algorithms in nonlinear optimization with a complete convergence analysis (certificate on number of iterations needed for given level of precision). The algorithm also gives convenient interpretation of the power algorithm, a very popular algorithm of numerical linear algebra.

The paper is featured in the Wikipedia entry of sparse PCA (http://en.wikipedia.org/wiki/Sparse_PCA). 100 citations on Google Scholar indicate applications in large-scale analysis of data ranging from genomics to astronomy.