We propose a novel convex relaxation of sparse principal subspace estimation based on the convex hull of rank-d d projection matrices (the Fantope). The convex ...

We propose a novel convex relaxation of sparse principal subspace estimation based on the convex hull of rank-d projection matrices (the Fantope).

We propose a novel convex relaxation of sparse principal subspace estimation based on the convex hull of rank-d projection matrices (the Fantope).

A novel convex relaxation of sparse principal subspace estimation based on the convex hull of rank-d projection matrices (the Fantope) is proposed and ...

Aug 8, 2013 · We propose a novel convex relaxation of sparse principal subspace estimation based on the convex hull of rank-d projection matrices (the Fantope) ...

Fantope Projection and Selection: Near-optimal convex relaxation of. Sparse ... and can be achieved by where the max is over s-sparse, rank-k projection matrices.

We propose a novel convex relaxation of sparse principal subspace estimation based on the convex hull of rank-d projection matrices (the Fantope).

The estimator is based on a convex relaxation of the sparse PCA problem based on the convex hull projection matrices (the Fantope).

Fantope Projection and Selection: A near-optimal convex relaxation of sparse PCA ... We propose a novel convex relaxation of sparse principal subspace estimation ...

Nov 7, 2013 · It can be estimated at optimal rate. • Near optimal practical method: Fantope + ADMM. • It works, but its behavior needs better understanding.