Non-parametric Bayesian Isotonic Calibration: Fighting Over-Confidence in Binary Classification.
Mari-Liis AllikiviMeelis KullPublished in: ECML/PKDD (2) (2019)
Keyphrases
- binary classification
- multi class
- support vector
- learning problems
- cost sensitive
- gaussian processes
- support vector machine
- multi label
- multi class classification
- prediction accuracy
- class imbalance
- binary classifiers
- generalization error
- ensemble methods
- bayesian networks
- active learning
- learning styles
- knn
- learning process
- feature selection