Deterministic Column Sampling for Low-Rank Matrix Approximation: Nyström vs. Incomplete Cholesky Decomposition.
Raajen PatelTom GoldsteinEva L. DyerAzalia MirhoseiniRichard G. BaraniukPublished in: SDM (2016)
Keyphrases
- low rank matrix approximation
- low rank approximation
- kernel matrix
- singular value decomposition
- low rank
- matrix completion
- subspace learning
- kernel methods
- data matrix
- missing data
- kernel function
- spectral clustering
- metric learning
- matrix approximation
- nonnegative matrix factorization
- reconstruction error
- data dependent
- feature space
- adjacency matrix
- incomplete data
- model selection
- missing values
- input space
- positive definite
- eigendecomposition
- ls svm
- latent semantic indexing
- training samples
- iterative algorithms
- dimensionality reduction
- reproducing kernel hilbert space
- matrix factorization
- data representation
- support vectors