Efficient Approximation of Cross-Validation for Kernel Methods using Bouligand Influence Function.
Yong LiuShali JiangShizhong LiaoPublished in: ICML (2014)
Keyphrases
- cross validation
- kernel methods
- support vector
- model selection
- kernel function
- reproducing kernel hilbert space
- error estimates
- training set
- hyperparameters
- generalization error
- machine learning
- kernel matrix
- support vector machine
- learning problems
- feature space
- ls svm
- kernel learning
- regularization parameter
- feature vectors
- loss function
- svm classifier
- kernel trick
- neural network