PAC-Like Upper Bounds for the Sample Complexity of Leave-one-Out Cross-Validation.
Sean B. HoldenPublished in: COLT (1996)
Keyphrases
- leave one out cross validation
- upper bound
- generalization error
- sample complexity
- lower bound
- vc dimension
- model selection
- sample size
- support vector machine
- worst case
- cross validation
- variable selection
- k nearest neighbour
- kernel learning
- feature selection
- confidence intervals
- feature subset
- high dimensional
- training set
- active learning
- concept classes
- simulated annealing
- nearest neighbor