Reducing Kernel Matrix Diagonal Dominance Using Semi-definite Programming.
Jaz S. KandolaThore GraepelJohn Shawe-TaylorPublished in: COLT (2003)
Keyphrases
- semi definite programming
- kernel matrix
- kernel methods
- kernel function
- kernel learning
- semidefinite programming
- metric learning
- low rank
- feature space
- model selection
- input space
- training samples
- support vectors
- kernel pca
- ls svm
- covariance matrix
- reproducing kernel hilbert space
- support vector
- convex optimization
- pairwise constraints
- semi supervised
- semidefinite
- eigendecomposition
- linear combination
- text classification
- maximum margin
- linear programming
- matrix factorization
- kernel trick