Some New Bounds on the Generalization Error of Combined Classifiers.
Vladimir KoltchinskiiDmitriy PanchenkoFernando LozanoPublished in: NIPS (2000)
Keyphrases
- generalization error
- generalization error bounds
- linear classifiers
- upper bound
- training set
- training data
- training error
- learning machines
- cross validation
- model selection
- training set size
- classification error
- generalization bounds
- uniform convergence
- learning algorithm
- rademacher complexity
- sample complexity
- boosting algorithms
- binary classification
- active learning
- algorithmic stability
- lower bound
- training and test sets
- vc dimension
- supervised learning
- sample size
- risk minimization
- subspace information criterion
- feature set
- test set
- target function
- classification accuracy
- weak classifiers
- conditional expectation
- support vector
- multi class
- np hard
- expected error
- risk bounds
- error rate
- machine learning algorithms
- svm classifier