Overly Optimistic Prediction Results on Imbalanced Data: Flaws and Benefits of Applying Over-sampling.
Gilles VandewieleIsabelle DehaeneGyörgy KovácsLucas SterckxOlivier JanssensFemke OngenaeFemke De BackereFilip De TurckKristien RoelensJohan DecruyenaereSofie Van HoeckeThomas DemeesterPublished in: CoRR (2020)
Keyphrases
- imbalanced data
- prediction accuracy
- ensemble methods
- sampling methods
- ensemble classifier
- class distribution
- linear regression
- feature selection
- support vector machine
- imbalanced class distribution
- classification models
- class imbalance
- highly imbalanced
- minority class
- imbalanced datasets
- decision trees
- ensemble learning
- high dimensionality
- test data