A Rayleigh-Chebyshev procedure for finding the smallest eigenvalues and associated eigenvectors of large sparse Hermitian matrices.
Christopher R. AndersonPublished in: J. Comput. Phys. (2010)
Keyphrases
- eigenvalues and eigenvectors
- singular value decomposition
- singular values
- correlation matrix
- covariance matrix
- eigendecomposition
- low rank approximation
- covariance matrices
- principal component analysis
- principal components
- coefficient matrix
- low rank
- low rank matrices
- symmetric matrices
- spectral clustering
- low rank matrix
- high dimensional
- sparse matrix
- adjacency matrix
- positive definite
- affinity matrix
- kernel matrix
- hessian matrix
- sparse coding
- pseudo inverse
- original data
- laplacian matrix
- sparse representation
- maximum likelihood
- feature extraction