Shifting, One-Inclusion Mistake Bounds and Tight Multiclass Expected Risk Bounds.
Benjamin I. P. RubinsteinPeter L. BartlettJ. Hyam RubinsteinPublished in: NIPS (2006)
Keyphrases
- multi class
- perceptron algorithm
- mistake bound
- multiclass problems
- lower bound
- support vector machine
- multiclass classification
- pairwise
- learning algorithm
- binary classification
- data dependent
- multi class classification
- upper bound
- cost sensitive
- feature selection
- multi task
- pac learning
- binary classifiers
- worst case
- semi supervised