(1 + epsilon)-class Classification: an Anomaly Detection Method for Highly Imbalanced or Incomplete Data Sets.
Maxim BorisyakArtem RyzhikovAndrey UstyuzhaninDenis DerkachFedor RatnikovOlga MineevaPublished in: J. Mach. Learn. Res. (2020)
Keyphrases
- detection method
- incomplete data sets
- class labels
- detection algorithm
- highly imbalanced
- classification accuracy
- pattern classification
- support vector data description
- feature vectors
- data sets
- class distribution
- incomplete data
- machine learning
- pattern recognition
- classification models
- support vector
- support vector machine svm
- text classification
- machine learning methods
- missing values
- feature extraction
- decision trees
- feature selection
- training samples
- unsupervised learning
- maximum likelihood
- high dimensional
- decision boundary
- imbalanced data