Diversity Sampling is an Implicit Regularization for Kernel Methods.
Michaël FanuelJoachim SchreursJohan A. K. SuykensPublished in: SIAM J. Math. Data Sci. (2021)
Keyphrases
- kernel methods
- reproducing kernel hilbert space
- kernel ridge regression
- kernel matrices
- kernel function
- reproducing kernel
- learning problems
- svm training
- kernel matrix
- machine learning
- support vector
- support vector machine
- feature space
- kernel machines
- tikhonov regularization
- learning tasks
- rademacher complexity
- random sampling
- multiple kernel learning
- multiple kernel
- kernel trick
- image processing
- high dimensional feature space
- graph kernels
- kernel pca
- high dimensional
- generalization bounds
- regularization parameter
- sample size
- semi supervised
- knn
- kernel based clustering