Sparse PCA via Covariance Thresholding

Yash Deshpande, Andrea Montanari

In sparse principal component analysis we are given noisy observations of a low-rank matrix of dimension $n\times p$ and seek to reconstruct it under additional sparsity assumption…