Exploiting sparsity in linear and nonlinear matrix inequalities via positive semidefinite matrix completion.
Sunyoung KimMasakazu KojimaMartin MevissenMakoto YamashitaPublished in: Math. Program. (2011)
Keyphrases
- positive semidefinite
- matrix completion
- low rank
- kernel matrix
- linear combination
- missing data
- convex optimization
- matrix factorization
- semidefinite programming
- singular value decomposition
- semidefinite
- high dimensional data
- convex relaxation
- semi supervised
- kernel function
- high order
- sufficient conditions
- singular values
- high dimensional
- kernel methods
- input space
- metric learning
- data sets
- multi label
- model selection
- linear programming
- active learning
- pairwise
- feature extraction