Sample Complexity and Performance Bounds for Non-Parametric Approximate Linear Programming.
Jason PazisRonald ParrPublished in: AAAI (2013)
Keyphrases
- sample complexity
- linear programming
- vc dimension
- upper bound
- upper and lower bounds
- lower bound
- covering numbers
- linear threshold
- theoretical analysis
- learning problems
- learning algorithm
- pac learning
- np hard
- special case
- average case
- active learning
- optimal solution
- uniform convergence
- training examples
- linear program
- generalization error
- constant factor
- concept classes
- sample size
- supervised learning
- lower and upper bounds
- worst case
- objective function
- statistical learning theory
- data sets
- generalization bounds
- linear functions
- number of irrelevant features
- irrelevant features
- machine learning
- mistake bound
- pac model
- learning tasks
- image classification
- small number