Surrogate losses and regret bounds for cost-sensitive classification with example-dependent costs.
Clayton ScottPublished in: ICML (2011)
Keyphrases
- regret bounds
- cost sensitive classification
- cost sensitive
- misclassification costs
- cost sensitive learning
- bregman divergences
- multi class
- class distribution
- online learning
- linear regression
- lower bound
- error rate
- naive bayes
- expected cost
- classification error
- binary classification
- missing values
- class imbalance
- upper bound
- active learning
- training examples
- test set
- e learning
- training samples
- long run
- binary classifiers
- cross validation
- linear predictors
- reinforcement learning
- decision trees