Existence and Minimax Theorems for Adversarial Surrogate Risks in Binary Classification.
Natalie S. FrankPublished in: CoRR (2022)
Keyphrases
- binary classification
- multi class
- neyman pearson
- support vector
- learning problems
- cost sensitive
- multi class classification
- support vector machine
- multi label
- generalization error
- prediction accuracy
- ensemble methods
- class imbalance
- binary classifiers
- minimax search
- learning algorithm
- cross validation
- worst case
- reinforcement learning
- neural network
- learning tasks
- class distribution
- pairwise
- lower bound