A regularized matrix factorization approach to induce structured sparse-low-rank solutions in the EEG inverse problem.
Jair Montoya-MartínezAntonio Artés-RodríguezMassimiliano PontilLars Kai HansenPublished in: EURASIP J. Adv. Signal Process. (2014)
Keyphrases
- matrix factorization
- low rank
- sparsity constraints
- low rank matrix
- rank minimization
- trace norm
- collaborative filtering
- low rank matrices
- nuclear norm
- tensor factorization
- missing data
- matrix completion
- factorization methods
- recommender systems
- regularized regression
- negative matrix factorization
- data matrix
- nonnegative matrix factorization
- kernel matrix
- linear combination
- singular values
- convex optimization
- low rank matrix factorization
- latent factors
- stochastic gradient descent
- high order
- high dimensional data
- factorization method
- sparse representation
- missing values