NI-MWMOTE: An improving noise-immunity majority weighted minority oversampling technique for imbalanced classification problems.
Jianan WeiHaisong HuangLiguo YaoYao HuQingsong FanDong HuangPublished in: Expert Syst. Appl. (2020)
Keyphrases
- noise immunity
- minority class
- class imbalance
- class distribution
- nearest neighbour
- classification error
- class imbalanced
- cost sensitive learning
- active learning
- retrieval accuracy
- imbalanced data
- decision boundary
- sampling methods
- feature selection
- concept drift
- cost sensitive
- image retrieval
- training set
- object recognition
- pattern recognition