Near-Optimal Bounds for Cross-Validation via Loss Stability.
Ravi KumarDaniel LokshtanovSergei VassilvitskiiAndrea VattaniPublished in: ICML (1) (2013)
Keyphrases
- cross validation
- model selection
- worst case bounds
- upper bound
- hyperparameters
- support vector
- error estimates
- cross validated
- training set
- lower bound
- variable selection
- generalization error
- regularization parameter
- regression problems
- learning machines
- data sets
- worst case
- ls svm
- error estimation
- regret bounds
- information criterion
- nearest neighbor classifiers
- feature selection