Regularization-free estimation in trace regression with symmetric positive semidefinite matrices.
Martin SlawskiPing LiMatthias HeinPublished in: CoRR (2015)
Keyphrases
- positive semidefinite matrices
- positive semidefinite
- complementarity problems
- polynomial kernels
- regression model
- reproducing kernel hilbert space
- convex sets
- embedding space
- kernel matrix
- semidefinite programming
- learning algorithm
- metric learning
- density estimation
- similarity function
- kernel methods
- shortest path
- least squares
- training data