Discrepancy and approximations for bounded VC-dimension.
Jirí MatousekEmo WelzlLorenz WernischPublished in: Comb. (1993)
Keyphrases
- vc dimension
- upper bound
- sample complexity
- lower bound
- sample size
- concept classes
- covering numbers
- inductive inference
- vapnik chervonenkis dimension
- statistical learning theory
- distribution free
- worst case
- empirical risk minimization
- pac learnability
- concept class
- generalization bounds
- compression scheme
- uniform convergence
- pac learning
- learning algorithm
- support vector machine
- multiresolution
- theoretical analysis
- supervised learning
- np hard