Discrepancy and epsilon-approximations for bounded VC-dimension
Jirí MatousekEmo WelzlLorenz WernischPublished in: FOCS (1991)
Keyphrases
- vc dimension
- sample complexity
- theoretical analysis
- upper bound
- lower bound
- covering numbers
- statistical learning theory
- learning problems
- concept classes
- pac learning
- special case
- vapnik chervonenkis dimension
- sample size
- learning algorithm
- distribution free
- generalization bounds
- generalization error
- active learning
- empirical risk minimization
- inductive inference
- target function
- pac learnability
- worst case
- supervised learning
- concept class
- compression scheme
- training examples
- uniform convergence
- uniform distribution
- kernel methods
- multiresolution
- linear functions
- function classes
- support vector
- image processing
- machine learning
- data sets