Lower Bounds for Training and Leave-One-Out Estimates of the Generalization Error.
Gérald GavinOlivier TeytaudPublished in: ICANN (2002)
Keyphrases
- generalization error
- upper bound
- lower bound
- cross validation
- training error
- training set
- training set size
- learning machines
- unbiased estimator
- test set
- supervised learning
- training and test sets
- classification error
- sample complexity
- model selection
- sample size
- target function
- learning algorithm
- vc dimension
- linear classifiers
- active learning
- confidence intervals
- worst case
- error estimation
- training data
- conditional expectation
- subspace information criterion
- training process
- np hard
- training samples
- support vector
- low error
- machine learning
- data sets
- real valued
- error rate
- training examples
- optimal solution
- objective function