Optimizing for ROC Curves on Class-Imbalanced Data by Training over a Family of Loss Functions.
Kelsey LiebermanShuai YuanSwarna Kamlam RavindranCarlo TomasiPublished in: CoRR (2024)
Keyphrases
- loss function
- roc curve
- receiver operating characteristic
- class imbalanced data
- learning to rank
- pairwise
- logistic regression
- support vector
- evaluation measures
- probability estimation
- feature space
- misclassification costs
- training samples
- training examples
- machine learning
- convex loss functions
- feature selection
- receiver operator characteristic