Active Learning with Abstaining Classifiers for Imbalanced Drifting Data Streams.
Lukasz KoryckiAlberto CanoBartosz KrawczykPublished in: IEEE BigData (2019)
Keyphrases
- roc analysis
- class distribution
- concept drift
- data streams
- active learning
- class imbalance
- imbalanced class distribution
- imbalanced datasets
- rare class
- misclassification costs
- data stream classification
- minority class
- streaming data
- cost sensitive
- imbalanced data sets
- imbalanced data
- training set
- sliding window
- training examples
- unlabeled data
- receiver operating characteristic
- ensemble classifier
- cost sensitive learning
- classification algorithm
- stream data
- change detection
- data distribution
- ensemble learning
- binary classification problems
- roc curve
- random sampling
- labeled data
- learning process
- data sets
- machine learning
- training data
- decision boundary
- binary classification
- sampling methods
- test set
- non stationary
- training samples
- rare events
- semi supervised
- supervised learning
- small number
- anytime classification
- breast cancer
- machine learning algorithms
- error rate
- feature selection