Almost Linear VC Dimension Bounds for Piecewise Polynomial Networks.
Peter L. BartlettVitaly MaiorovRon MeirPublished in: NIPS (1998)
Keyphrases
- vc dimension
- piecewise polynomial
- vapnik chervonenkis dimension
- upper bound
- vapnik chervonenkis
- lower bound
- sample size
- covering numbers
- distribution free
- sample complexity
- generalization bounds
- statistical learning theory
- empirical risk minimization
- concept classes
- inductive inference
- uniform convergence
- learning machines
- worst case
- concept class
- pac learning
- function classes
- learning algorithm
- training data
- compression scheme
- b spline
- risk bounds
- upper and lower bounds
- learning rate
- learning theory
- learning problems
- model selection
- data points
- learning process