Undersampling is a Minimax Optimal Robustness Intervention in Nonparametric Classification.
Niladri S. ChatterjiSaminul HaqueTatsunori HashimotoPublished in: CoRR (2022)
Keyphrases
- classification accuracy
- feature extraction
- class imbalance
- worst case
- image classification
- automatic classification
- machine learning methods
- pattern recognition
- binary decision tree
- classification method
- dynamic programming
- support vector machine svm
- density estimation
- preprocessing
- benchmark datasets
- unsupervised learning
- text classification
- decision trees
- feature space
- supervised learning
- support vector machine
- multi class
- high dimensionality
- classification models
- knn
- optimal solution
- classification scheme
- feature vectors