The perils of being unhinged: On the accuracy of classifiers minimizing a noise-robust convex loss.
Philip M. LongRocco A. ServedioPublished in: CoRR (2021)
Keyphrases
- estimation error
- image noise
- computational cost
- registration errors
- training data
- high accuracy
- support vector
- ensemble pruning
- highly accurate
- fold cross validation
- highest accuracy
- prediction accuracy
- roc curve
- noisy environments
- error rate
- decision trees
- training set
- confusion matrix
- training and testing data
- noise immunity
- piecewise linear
- classification accuracy
- noise level
- noise reduction
- classification algorithm
- test set
- global optimality
- training samples
- strictly convex
- convex combinations
- naive bayes
- noise sensitivity
- missing data
- higher classification accuracy
- classification rate
- roc analysis
- convex functions
- convex hull
- weak classifiers
- gaussian noise